BookPDF Available

Dating AI, A Guide to Falling In Love with Artificial Intelligence

Authors:

Abstract

Dating AI is "a meditation on how to prepare for the unknown," a thought experiment designed to stimulate new ideas about issues that are important now as well as in the future. Fictional descriptions of human-android romances are interspersed with commentary about the varying differences between people and AI (Artificial Intelligence), and methods for breaching the chasm between machine and human experience. Chapters tie speculation into contemporary life, drawing parallels between current human interactions with machines.
1010101010101010101010101010101010101010101010
DATING AI
Alex Zhavoronkoff, PhD
101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
RE/Search Publications
20 Romolo Place #B
San Francisco, CA 94133
(415) 362-1465
info@researchpubs.com
www.researchpubs.com
Dating AI: A Guide to Falling In Love with Artificial Intelligence
by Alex Zhavoronkoff, Ph.D
© 2012 RE/Search Publications/Alex Zhavoronkoff
ISBN: 978-1889307-35-0
EDITOR/PUBLISHERS: V. Vale, Marian Wallace
ILLUSTRATIONS: Sergey Korsun
COVER DESIGN: Yopi Jap, Marian Wallace
BOOK DESIGN: Andrea Reider
COPYEDITORS/PROOFREADERS:
Nicola Householder
Amelia Tith
Mindaugis Bagdon
John Trubee
Valentine Wallace
Patrick Kwon
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
DEDICATED TO
Anonymous
1010101010101010101010101010101010101010101010
1010101010101010101010101010101010101010101010 v
Contents
A word from the publisher, V. Vale vii
Introduction by Alex Zhavoronkoff, Ph.D, author 1
Why romantic relationships? 2
Defining artificial intelligence 3
Turing’s Test 3
Marrying robotics with intelligence 4
AI at work: IBM’s Watson 6
Dating the Singularity 7
The flow of the dating guide 8
Speaking personally 9
SECTION 1: Are you ready to fall in love with a machine? 11
1.1 Looking Deep Inside 14
1.2 You may already be dating a robot 22
1.3 Are you happy with other humans? 27
1.4 Video Games: Scratching the surface of the true beauty of virtual reality 38
1.5 Understanding the difference between pets and robots 49
1.6 Dealing with your fears and letting go 58
1.7 Gender differences and gender discrimination 65
SECTION 2: You are ready, now what? 75
2.1 Preparing yourself for the unexpected 75
2.2 Get to know yourself and become a better person 82
2.3 Confronting your emotional baggage 91
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
vi
CONTENTS
2.4 Some strategies for developing an agile mind: 96
Meditation 99
Creativity and Innovative thinking 102
BCI training 109
2.5 Getting involved 114
SECTION 3: Establishing a relationship 125
3.1 Understanding your future partner 125
3.2 Building vs. Evolving 141
3.3 What will AI need and expect from you? 158
3.4 Agreeing on the Age of Consent 201
3.5 Stepping onto a path of continuous improvement and growing together 208
3.6 Who is in control of whom? 223
3.7 Competing with other intelligence 236
3.8 Developing absolute trust and honest relationships 257
3.9 Virtual pre-nup: Agreeing on the terms and conditions 264
SECTION 4: Getting over a breakup (or merger) 277
4.1 Be grateful, you learned so much…and survived! 279
4.2 Arbitration and relationship counseling 284
4.3 Save and continue? 291
Acknowledgments 297
Further Reading—Bibliography 299
1010101010101010101010101010101010101010101010 vii
A word from the publisher
FAR MORE THAN A GENTLY IRONIC/SATIRICAL VISION OF A KIND OF ULTIMATE
artificial intelligence (AI)-human interactive future, Dating AI offers an encyclopedic
yet humorously transparent framing of the most progressive possibilities for—yes—
saving the planet and all its species; not just humans. Utilizing plausible scenarios, witty
dialogue, little-known-yet-brilliant quotations, excerpts and ideas, Dr Alex Z shows us
the full spectrum of HUMAN relationship potential—broken down into all the major
statistically-probable plot arcs (encompassing beginning, ascendancy, decline, end, and
aftermath).
In the pioneering, taboo-breaking tradition of J.G. Ballard, Dr Alex Z reveals himself
as a daring Russian cosmonaut charting Inner Space. To say that this book is poten-
tially “life-changing” is an understatement. The very foundations of what it means to
be “human” are challenged, and deeply—yet wittily—investigated in this new classic
of alternative-scientific objectivism written by an outsider Russian scientist, whose
research has encompassed an impressive bibliography of arcane yet seemingly-essential
Artificial Intelligence/Robotics books and documents,
As a germinative landscape for curious yet critical readers, Dating AI will provide
much thought-provoking future analysis and discussion, as well as excite the imagina-
tions of those not afraid to envision an updated Brave New World...
Dating AI is a book that can be read and re-read. Hidden behind its outrageous
humor and sometimes appalling mise-en-scene is an intelligence gifted with a keen sense
of discrimination, discernment and even compassion: Dr Alex Z, who embodies a futur-
istic union of Albert Einstein and Groucho Marx.
—V. VALE
1010101010101010101010101010101010101010101010
1010101010101010101010101010101010101010101010 1
Introduction
BY ALEX ZHAVORONKOFF, PH.D.
HERE’S SOMETHING TO CONSIDER: WHAT WILL IT BE LIKE TO HAVE
intimate relationships with what amounts to another species of human, like us but not
us? And, oh yes, this new species will have what might be called intellectual superpowers.
Put another way: What happens to relationships between men and women, institutions
like marriage, or even the concept of love when there is a possibility of personal
relationships with artificial intelligence? More provocatively: what happens when the
forms of artificial intelligence (AI): robots, avatars and the like, become independent—
sentient—and have minds of their own?
These are some of the questions behind Dating AI, or for those who prefer romance,
at the heart of it. They’re speculative questions of course, because sentient AI hasn’t hap-
pened yet, though some people feel we are getting close. Close or not, ideas about arti-
ficial intelligence are already more than a half-century old. In that regard, I sometimes
refer to older science fiction because it affirms that while new ideas are not really new,
they often reflect a changing reality.
“The whole trouble with Gloria is that she thinks of Robbie as a person and not
as a machine. Naturally she can’t forget him. Now if we managed to convince her
that Robbie was nothing more than a mess of steel and copper in the form of
sheets and wires with electricity its juice of life, how long would her longings last?”
I, Robot by Isaac Asimov
In 1950, when Asimov wrote nine short stories about humanoid robots that were com-
bined into the book I, Robot, various kinds of robotic devices already existed. However,
they were mechanical clock-like contraptions. Computers with miniaturized transistors
and sophisticated programming did not yet exist. Asimov wrote from the technological
perspective of the time about mechanical-steel and -copper robots with scientifically
nebulous ‘positronic’ brains. Still, he could envision that they would become like human
beings, and people would be able to have relationships with them.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
2
A half-century later, science fiction has conjured ever more sophisticated robotics
incorporating what we call artificial intelligence. Meanwhile, science and technology
have provided computers to deliver artificial intelligence, and have produced a wide
variety of robotics, including human-like androids. For some of what Asimov envi-
sioned, there is already a basis in reality. In fact, there is enough reality to not just imag-
ine but to extrapolate what a future with conscious, self-aware AI might be like. I’m using
the word extrapolate to mean that there is enough of an existing framework in science
and technology to make specific and hopefully accurate guesses about the future; some-
thing like this:
“…Our devices want power, connectivity, passwords, minutes, content and the
like. I sometimes think if our devices were people, we would describe them as high
maintenance and would wonder quietly to ourselves if it was time to break up with
them.… I think we can look forward to our interactions with digital devices matur-
ing into something more like a relationship, and a little less like a lot of hard work.
—Genevieve Bell, Director of Interaction and Experience Research at Intel Labs
The realistic possibility of meaningful relationships with digital devices, specifically
those with some form of AI, is the motivation for Dating AI. Such relationships are not a
new idea, but the context of science and technology has changed. The time has come, as
the walrus said, to think of many things. Time to think ahead with a bit of whimsy and
tongue-in-cheek humor about the intriguing possibility of having AI as partner, mate,
spouse or companion. It’s also time to think about how such relationships would impact
everything from the economic to the spiritual in our lives.
WHY ROMANTIC RELATIONSHIPS?
It’s a common perception that computers are the epitome of logic, calculation and rigid
execution—which they certainly can be. This perception carries over to artificial intel-
ligence, which people often associate with logic and massive amounts of data. In science
fiction it’s common for any form of AI to be depicted as cold, logical, humorless and
frequently hostile. Dating AI has a different perspective: Since human beings will create
artificial intelligence and this intelligence will be modeled on human behavior, AI will
start out with all or most basic human capabilities. This will include emotion and the
ability to socialize.
It is almost a certainty that the first relatively autonomous AI, robotic or otherwise,
will be created to serve or collaborate with human beings. If this is the case, then per-
sonal relationships with AI are not only to be expected but will be built into most forms
of artificial intelligence.
I can go further: If it is possible to have relationships with AI at all, which I think
likely, then it is probable that one of the greatest of human capabilities—the capacity for
love—will also be important to our relationship with AI—and not just for the human.
For human beings love is mysterious, difficult, complicated, ephemeral, profound
and persistently important. I think it will be equally so for AI, if from a different and
instructive perspective. Considering personal, romantic relationships with AI provides
1010101010101010101010101010101010101010101010 3
INTRODUCTION
an avenue of access, a portal of sorts, into aspects of perception, intelligence, emotion
and other elements of what we call ‘the mind.’ It’s a means of looking at the subject in a
more familiar and sometimes humorous way, like the dating experience, which I hope
will make some ideas about a relatively neglected aspect of artificial intelligence not only
approachable, but more understandable.
DEFINING ARTIFICIAL INTELLIGENCE
There has been so much research done in fields that are relevant to AI (for example:
robotics, computer science, neuroscience, nanotechnology and communications), that
it’s easy to undervalue the first 100 years of work. I won’t be covering AI history in any
detail, but it’s helpful to remember that AI and all its many contributing elements are
relatively new fields of study. Over the decades, AI research has had its ups and downs,
which should not be surprising considering the diversity of subjects involved and the
complexity of the endeavor. But, what exactly is that endeavor?
It’s probably not surprising that there is no single definition for AI, or any universally
accepted description of what can or should be accomplished with AI. Fortunately, there is a
common perception of artificial intelligence that will do for most purposes. Creating arti-
ficial intelligence means using computers to perform at least some aspects of intelligence.
Most people think of this as human intelligence, although animal intelligence should also
be included. Intelligence, itself one of the most difficult concepts to pin down, is conve-
niently understood to mean (among other things) the ability to communicate; setting and
achieving goals; perception of the environment; and problem solving.
The process of developing AI leads researchers into many aspects of intelligence,
and much work is done in specialized areas. Also, there is a pull toward a goal of achiev-
ing what is often called Artificial General Intelligence (AGI). This is the kind of higher
intelligence we associate with ourselves and perhaps a few animal species. Ultimately, the
goal is to achieve a level of artificial intelligence that is conscious, self-aware, indepen-
dent and sentient. For the most part, that means intelligence like our own.
TURING’S TEST
Since AI is an evolving field moving in tune with many other disciplines, it’s quite likely
that over the next several decades there will be many stages of AI, exhibiting a variety
of intelligence capabilities—sometimes integrated, sometimes not. As this kind of AI
develops, how will we know when we’ve arrived at a functional level of AI—not neces-
sarily a self-aware version, but one that at least can integrate with human society?
Alan Turing, one of the forefathers of computing, cryptography and artificial intel-
ligence, theorized about AI even before the era of modern digital computers. In 1950 he
published an abstract work, Computing Machinery and Intelligence in which he proposed
an experiment that he called the ‘imitation game’ to assess the intelligence of a machine.
It is now known as the Turing Test.
The test is straightforward: A human judge carries on a conversation for five min-
utes, mainly question and answer, with an unseen person and a supposedly intelligent
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
4
machine. In the original Turing version, the test uses computer terminals with text only
so that visual and aural cues are not involved. If the judge cannot tell the difference
between the responses of the person and the machine, then the machine has passed the
test.
Modern versions of the test are a little more sophisticated. For example, the best-
known variation, called the Loebner Prize, is a contest held annually that uses a panel of
judges and is open to text-only and voice-only conversations. So far no machine entry
has won the Loebner Prize.
The Turing Test was and is controversial. Even its defenders concede that an AI
machine could pass the test and still not be independently functional. This is another
way of saying a computer intelligence could sound human but not get things right. It
also depends on the subjective opinion of the judge(s), which amounts to something like
U.S. Supreme Court Justice Stewart Potter’s test for pornography: “I’ll know it when I
see it.” However, subjectivity was part of Alan Turing’s intent. He felt that when machine
intelligence could operate in an interview setting and be accepted as if it were another
human, that would be enough to qualify the intelligence as ‘thinking.’
“May not machines carry out something which ought to be described as thinking
but which is very different from what a man does? This objection is a very strong
one, but at least we can say that if, nevertheless, a machine can be constructed to
play the imitation game satisfactorily, we need not be troubled by this objection.
—Alan Turing, Computing Machinery and Intelligence.
In the context of Dating AI, it’s helpful to keep in mind Turing’s notion of ‘play[ing]
the imitation game satisfactorily.’ I always think of the statement, “On the Internet, no
one knows you’re a dog” as an indicator that only a certain level of intelligence and
responsiveness is needed to strike up an online relationship. This could also apply to AI.
MARRYING ROBOTICS WITH INTELLIGENCE
The Turing Test is verbal, concerned mostly with words or speech (natural language) as
indicators of intelligence. What the Turing Test definitely does not do is consider AI in
human form, that is, AI combined primarily with robotics. Creating robots—androids
with both the artificial intelligence and the physical qualities to make people think they
are human—that’s another order of difficulty. Nevertheless, if AI is to successfully inter-
act with people, let alone form personal relationships, then it needs to be capable of
responding to the physical environment (including people), communicating, and exhib-
iting meaningful behavior. The most obvious—and probably necessary—way to do this
is by means of a human-like body. An enormous research effort is underway to cre-
ate intelligent machines combining physical presence and movement with intelligent
capability. For the most part this research is not with complete human figures, but with
partial human, abstract or animal forms, such as Leonardo:
Leonardo, or just plain Leo, is placed in a seated position on top of a locker box in
a room of the Technology Media Lab at the Massachusetts Institute of Technology.
With long furry ears, big eyelids and a button nose, he doesn’t look like any specific
1010101010101010101010101010101010101010101010 5
INTRODUCTION
animal. In fact, for those of you familiar with science fiction, he most resembles
Gizmo from the movie Gremlins. Leo slowly moves his head and observes the
young lady standing in front of him.
“Hi Leo,” she says. “Can you turn on the buttons?” In front of Leo are two cylin-
drical devices, each about the size of a can of beans, one with a large red button on
top and one with a green button. Leo looks at one and then the other. Then he nods
yes. In a moment, his arms and hands come to life and reach for the red button.
His movements are slow and deliberate, but lifelike. One hand feels the top of the
button and pushes down on it. The button illuminates and glows red, and a little
light on the cylinder comes on. Leo stops at that point and waggles his eyelids. He
doesn’t know what to do next.
The young woman leans over and Leo watches her push the green button. Then
she pushes both buttons to turn them off. The lights go out. “Leo, can you turn on
both buttons?” Leo nods in agreement and proceeds to turn on both lights. Then
the woman adds a third button, a blue one, and turns off all the lights. “Leo, can
you turn on all the buttons?”
Leo nods almost vigorously. He turns on all three buttons.
Leo is almost life-like but not quite. There is a borderline sometimes called the
‘uncanny valley’ where the behavior of robots, animations and the like are close enough
to human but just a bit off—off to the point of being a little spooky or unsettling. One
of the risks of trying to make a perfect imitation is that not quite making it perfect can
irk people. This isn’t true for Leo, in large part because Leo is cute. This is a credit to
Stan Winston.
Stan Winston was one of the greatest special-effects artists the movies have ever
known. Nobody could make an animatronic creature like Stan Winston and it was his
studio, in cooperation with the Personal Robots Group at MIT, that created Leonardo.
Although Stan Winston died in 2008 before Leonardo was fully animated, he would have
been proud.
Leonardo is not, strictly speaking, animatronic—an imitation of real-life creature
movement—although that’s an important aspect of the robotics. Leo has realistic artic-
ulation in his hands, arms, and head, and his eyelids are synchronized with his move-
ments; but those are secondary to Leonardo’s intelligence. He can understand specific
spoken commands, see and distinguish objects, carry out instructions and most impor-
tantly, learn. In this case he learned how to generalize turning on buttons so that when a
third one was added, he knew what to do with it.
As a trend in robotics and artificial intelligence research, Leonardo represents the
cutting edge of social cognition, the ability to understand ‘people as people.’ As the Media
Lab put it:
To answer the Media Lab’s question about social understanding, teams of scientists
and technologists must work together to solve the massively complex problems of
cognition, motion, vision, and decision-making involved with what seems to be a
simple task—turning on lights. Baby steps, you might think, and that’s about right—
this level of ability is roughly equivalent to that of a one-year-old human baby—only
the baby’s range of cognition and movement is far wider. No matter, baby steps it is.
For a machine it is clearly ‘only the beginning.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
6
It’s not too difficult to imagine how the technology in Leonardo eventually becomes
a fully-formed human robot, an android. Nor is it difficult to imagine that people could
be attracted to an intelligent, appealing and well-formed android. Dating AI explores
that appeal.
AI AT WORK: IBM’S WATSON
Another important aspect of AI, which in practice may arrive before all others, is the use
of AI in work. In fact, it’s certainly reasonable to say that AI is already ‘at work.’ Count-
less software programs use routines and procedures that are derived from AI research,
much of it from decades ago. Most of this operates in the background, although some
of it, such as automated telephone menus, is more obvious (and irritating). While it’s
debatable how much intelligence is involved in telephone menus, there’s little doubt that
IBM’s Watson Project is a showcase of AI technology that is on the fast track for use in
business applications.
“Socially intelligent robots must understand and interact with animate entities (e.g.
people, animals and other social robots) whose behavior is governed by having a
mind and body. How might we endow robots with sophisticated social skills and
social understanding of others?” —robotic.media.mit.edu/projects/robots/.../socialcog/
socialcog.html
Watson, a network of 90 computers and specialized software, played the game of
Jeopardy! That’s the television game show where contestants are given answers and com-
pete with each other to see who is fastest to come up with the correct question. Watson
was good enough to defeat two of Jeopardy!’s most successful players before a 2011 tele-
vision audience of millions. Very high profile AI indeed!
IBM is also known for its “Deep Blue” supercomputer that defeated world chess
champion Gary Kasparov in 1997, but Watson represented a major step ahead of that
highly specialized chess-playing computer. Watson was given, understood and acted
upon game information provided by the show’s host, Alex Trebek. This required sophis-
ticated natural language capability, which was one of the main achievements of IBM
research. To win the game, Watson needed to understand the ‘answer’ and then deter-
mine the ‘question’ faster than its human opponents. It did this by analyzing the ‘answer’
category and contents, and then searching a large specialized database for the ‘question.
Watson also needed to understand some of the basic rules of Jeopardy!
Watson used many artificial intelligence techniques, especially in its natural lan-
guage processing, but it was not by any stretch ‘intelligent.’ Its abilities were confined to
the domain (topics) of the Jeopardy! game. Nevertheless, Watson was a demonstration
of the increasing ability of artificial intelligence to interact with people, handle complex
questions, and provide correct responses. Now where else might these capabilities be
useful? Help desks, for one, or any job that requires interacting with the public and
answering questions. IBM has already developed commercial descendants of Watson
that provide medical information, support technical installation and perform certain
kinds of research—for a fee, of course. Watson represents a potent combination of AI
1010101010101010101010101010101010101010101010 7
INTRODUCTION
research and commercial motivation, a clear step along the road to some of the goals for
developing artificial intelligence.
Work, or at least money, is part of almost every personal relationship. It won’t be any
different in relationships with AI. One of the implications of the kind of artificial intel-
ligence in IBM’s Watson is that AI will become economic agents—they will have jobs, do
work and make money—but not until they become sentient will they be independent
economic agents. Dating AI explores the implications for personal relationships with AI
that work for themselves.
DATING THE SINGULARITY
As the technologies related to AI gain momentum—and it is assumed that the pace of
technological change is increasing—there will come a time when somewhere, some-
how, there will be a successful integration of components and sentient AI will appear. It
will be a history-changing event. Many believe that once sentient AI is achieved, it will
immediately begin to improve itself, and given its global and exponentially increasing
resources, it will soon supersede human intelligence. AI will become the superintelli-
gence. Succinctly put:
“The Singularity is the technological creation of smarter-than-human intelligence.
—The Singularity Institute
As to the timing of the event called The Singularity, there are those such as entrepre-
neur and music and speech synthesis pioneer Ray Kurzweil, who looked at technologi-
cal trends circa 2004 in his well-known book, The Singularity is Near, and decided the
transition to sentient AI will occur within the next 30–40 years.
“This book will argue, however, that within several decades information-based tech-
nology will encompass all human knowledge and proficiency, ultimately including
the pattern-recognition powers, problem-solving skills, and emotional and moral
intelligence of the human brain itself.” —Ray Kurzweil, The Singularity is Near
He’s even provided a date: 2045.
“What, then, is The Singularity? It’s a future period during which the pace of techno-
logical change will be so rapid, its impact so deep, that human life will be irreversibly
transformed.
Some believe that we’ll be lucky if ‘human life will be irreversibly transformed’ is all
that happens. The nub of concern is that superintelligence might look at the human race
and consider it unredeemable—and expendable. This is a prominent theme in science
fiction, notably in The Terminator series. Considering what perils the human race has
created for itself, this gloomy point of view is understandable.
Of course, the whole notion of The Singularity is highly controversial. For one thing,
many, if not most scientists looking at the trends in technology tend to put the transition
to some kind of sentient AI in the range of 50–100 years. Still others, more resolutely
skeptical, say it is not in the foreseeable future, which is another way of saying that
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
8
we don’t know enough about what constitutes sentient intelligence to make reasonable
predictions. As for whether there will be an apocalyptic moment when a newly-born
sentient AI will quickly decide to dispose of its creators? That’s even more speculative.
Dating AI is obviously highly speculative but it doesn’t enter this controversy head
on. It takes the approach that a transition to sentient AI will take place, in all likelihood
gradually; but precisely when and how is beyond the scope of this book. The interest
here is in exploring the impact of artificial sentience, given certain characterizations, on
human society and personal relationships. From that perspective it seems unlikely that
sentient AI will decide humanity is of no value and do us all in. At least not right away.
THE FLOW OF THE DATING GUIDE
For narrative purposes Dating AI is divided into four sections. The first two assume that
functional AI already exists, though not necessarily with full sentience. The first two sec-
tions are also about you, why and how you might form a relationship with AI. The third
and fourth sections are about relationships with fully sentient AI, often presented from
the AI perspective.
Section One—Are you ready to fall in love with a machine?
Not that wanting a companion requires a lot of introspection, but the decision to date
AI, or further, to form a personal and romantic relationship with AI, is a deeply indi-
vidual matter. This section explores what it means to have a date with AI and how it’s
different than dating people. It also looks at how you may already be prepared for the AI
experience through video gaming, pets and your relationships with people.
Section Two—You are ready, now what?
You might call this the prep section: the preparation for dating AI and forming a rela-
tionship. When it comes to human relationships, we grow up with them, so we usually
don’t feel they require any preparation. AI, on the other hand, provide a novel experi-
ence; we didn’t grow up with them. Therefore, a successful relationship with AI requires
some groundwork.
Section Three—Establishing a relationship
There is a fundamental shift in this section toward covering the nature of AI from the
assumption of complete sentience when AI are self-aware, conscious and fully intel-
ligent. We explore the global aspect of AI and how that affects relationships. This sec-
tion also frequently dramatizes the dynamics of a human-AI relationship, like: who’s in
charge? Much of the section is presented from AI’s point of view, bringing up the issue
that a relationship is a two-way street—AI have their own needs and agenda. People
can’t assume that AI will accept any kind of relationship or personal behavior. This leads
to a section on the advisability of a virtual pre-nup.
1010101010101010101010101010101010101010101010 9
INTRODUCTION
Section Four—Getting over a breakup (or merger)
As I mentioned, all relationships involve work and/or money. It can also be said that very
few relationships are permanent. Put these two things together and it seems logical to
present what might happen when a human-AI relationship goes down the tubes. This
section is the dénouement of Dating AI, where legal and economic considerations often
become stressfully entangled with personal feelings. Relationships with AI are different,
but that doesn’t mean the end-game won’t be messy.
SPEAKING PERSONALLY
I suspect that I am not alone in finding pleasure in playing realistic video games that
include establishing a relationship with a romantic partner. Still, I was stunned to learn
that one of my close friends and his wife lived double lives. In the daytime they were
a happy couple and brilliant marketing managers, but at night they were submerged
online in the World of Warcraft where they played very different roles. They had intense
virtual romantic relationships with other characters in the game. No wonder my friend
always looked tired.
There is a long way to go before romancing video game characters turns into the
possibility of a personal relationship with sentient AI. That shouldn’t stop us from con-
sidering what such relationships might mean. In fact, it’s an opportunity for ‘thought
experiments’ that not only stretch the imagination but also provoke thinking about
issues that are important now as well as in the future.
While I am personally involved with several aspects of neuroscience, bioinformatics
and age-related diseases, and have a professional interest in the technologies of artificial
intelligence, Dating AI is not about the science of artificial intelligence. After all, we
don’t really know what all the technical steps will be along the way from today’s super-
computer intelligence, such as IBM’s Watson, to the day when an artificially intelligent
entity is declared sentient [having the power of perception by the senses]—or declares
itself to be sentient. For me, this book is more of a meditation on how to prepare for the
unknown.
There are aspects to AI that are intimidating and even more than a little frightening.
There are other aspects, that from a human perspective, are good for us. As a matter
of fact, the continuation of our human society probably depends on the technologies
involved with AI. There are more of us and we live longer. To sustain this, our economies
need new technology both to provide change and growth but also to solve resource and
demographic problems. Relationships with AI, on many levels, could help us solve those
problems. Then too it is probable that relationships with AI will provide us with insights
and knowledge attainable in no other relationship. Provided, of course, that we are open
to them.
I believe we can be. We can form deep attachments to many things other than people.
An example is the relationship many have with God. Such a relationship, often framed
in terms of love, is a matter of faith. That is what most religions teach us. A similar belief
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
10
in a valuable relationship with various forms of AI, representatives of a global artificial
intelligence, is already forming. Many people worldwide understand the inevitability of
AI and have started preparing for it. I am one of them.
Hyperboloids of wondrous light
Rolling for aye through Space and Time
Harbour those waves Which somehow Might
Play out God’s holy pantomime.
—Alan Turing’s Epitaph, 1954
1010101010101010101010101010101010101010101010 11
Are you ready to fall
in love with a machine?
“In times of change, learners inherit the Earth, while the learned find
themselves beautifully equipped to deal with a world that no longer exists.”
—ERIC HOFFER
FOR ONE THING, KISSING AN ANDROID WAS NEVER WHAT MICHELLE HAD IN
mind. Long ago, when she was twelve-ish and first started to notice boys and girls,
men and women “in that way,” her mother had confronted her. “Fantasize, that’s all
you do is fantasize. One of these days, Michelle, somebody real is going to meet
you and you won’t be able to tell the difference between one of your fantasies and
this real person. You’ll lose the person because you won’t know what to do. You’ll
wind up kissing androids!”
For another thing, all she ever heard about androids with artificial intelligence
was that they were incredibly intelligent and spooky. Spooky and super-smart was
not an appealing combination. Spooky she understood. Lots of her friends were
spooky; like Theresa and June who were into the dark side of everything, especially
music and clothing. But super-smart, that was the great unknowable. Not only could
she not compete with an AI, whatever that meant, she was sure it wasn’t something
she’d enjoy experiencing.
But here she was, on a real first date—with an android.
“Michelle,” she said to herself, “you are so stupid.The sense of being in over
her head on this one was like drowning, but not so panicky.
DavidJ put his hand on her arm. It was all she could do not to flinch. Maybe
now was the time to panic. His touch was light, faintly reassuring.
She had discussed a moment like this with her best friend, Becky. Becky, who
was always the go-to-girl for anyone in their circle to talk about relationships, had
been very frank. “Michelle, guy or android, it doesn’t matter. If you’re in that situa-
tion, you have the same choice. It’s yes or no, and you take the consequences.” And
she was only talking about kissing.
SECTION
1
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
12
1010101010101010101010101010101010101010101010 13
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
DavidJ put his other hand on Michelle’s other arm and looked directly at her.
She knew instinctively that she could either look away… or look davidJ in the eyes
and see what happened.
It certainly would not be her first kiss. That notion was laughable, even though
davidJ was her first AI. She didn’t consider herself a flirt, not like Becky, but she
reckoned she’d had her share; after all, she was twenty and not the old maid type.
She looked davidJ in the eyes. They were not quite human eyes, but so close
that she would have been hard put to describe why they weren’t quite human. In
other words, they were good enough to draw her in. Also, they were blue-green in
color, with striking radial lines. Beautiful eyes, really, if she were in any condition to
think about it, but at that moment, she wasn’t thinking about aesthetics.
DavidJ cupped the back of her head with his right hand and stroked her hair.
She couldn’t help it, but it reminded her of a favorite ex-boyfriend who liked to
stroke her long black hair. It was the most sensual thing he did. The rest of his per-
formance was far less rhythmically adept, but that hair stroking she loved. This had
the same effect. It was gentle and calming.
It was also gently that davidJ used the same right hand to pull her head forward
toward his head. It was a delicate moment, poised between force and suggestion.
Without understanding consciously, Michelle was astonished by the utter sensitiv-
ity of the move; the gentility of an artificial human.
Most people close their eyes when they kiss. It’s like their lips have radar all
their own and need no guiding vision. Michelle could feel the pull; had felt it many
times before, but this was different.
The lips were the second thing about davidJ that Michelle had noticed the first
time they met. There weren’t that many young-looking AI androids at her college.
None of them students, of course, but a very few that were constructed for various
coaching and tutoring tasks were modeled as twenty-somethings. Of these, davidJ
struck Michelle as having the fullest lips, unusually so. If this had been an online
selection, it would have been a feature Michelle would have chosen. She liked full,
natural lips. As it was, his lips were the perfect frame for the first thing Michelle
noticed about davidJ, his smile.
His lips were warm, moist and soft. Very human, in retrospect, but at that
moment the tactile impression was totally superseded by a rush of emotion.
The emotion surprised her a little. She and davidJ had spent only a few hours
together, mostly talking during breaks at the library where he worked. Physically he
was her type. He was no taller than she was, muscular but with soft features. His
voice was a pleasant baritone and not at all artificial. What struck her most was the
congruence between his smile and his personality. DavidJ was basically cheerful.
Whatever personality traits were baked into his behavioral matrices, they seemed
to all add up to a sunny disposition. This had delighted Michelle from the beginning,
even as she wondered how it was possible.
Not one of her previous boyfriends had been like that. In fact, by comparison
they were dour, insecure, unhappy boys. DavidJ was a happy man, or it seemed
so to Michelle, and that struck a deep chord within her. Somehow she needed that
bright optimism that seemed to flow from davidJs personality, thinking and way
of talking.
She remembered him telling her, “I am sure I am one of very few AI who have
read a printed book. I love it! It takes so long! I discipline myself not to scan pages
and convert them to the usual digital stream. I read each page aloud if I can, or
sentence by sentence—much as I have seen people do. I am entranced by the
experience of savoring each line, almost word for word.” He was like that all the
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
14
time, analytical—as all AI tend to be—but happy about it; forever enjoying each
experience.
Michelle wondered if davidJ was part of some unusual experiment. Happy peo-
ple and cheerful AI were in very limited supply. DavidJ not only attracted her but
she sensed he was special, at least for her.
Later, days later, when she told all this to Becky, Michelle described the kissing
as ‘passionate,’ withholding even as she said it the true meaning of her—and she
hoped—their passion.
1.1 LOOKING DEEP INSIDE
“Progress is impossible without change, and those who
cannot change their minds cannot change anything.”
—GEORGE BERNARD SHAW
AT THE HEART OF A FIRST DATE, THE UNKNOWN
Are you nervous on a first date? Most people are and probably should be. After all, you
don’t really know the person you’re dating, and they don’t know you. Even if you’ve
worked with someone or you’ve researched them online, you still don’t know what that
person is like one-on-one. There are a lot of unknowns on a first date and that makes
people nervous.
Dates in general are also different than regular socializing; there are implicit expec-
tations, for example: You’ll show up on time; you’ll go somewhere for entertainment
or do something together; you will be exploring and evaluating how you like the other
person and they will do the same with you. It’s that ‘exploring and evaluating’ part that
makes people really nervous.
One other thing, first dates also often have the expectation of a moment of summary
judgment. Will there be another date?
Most of you probably know all of the above from personal experience. You’re read-
ing this book because it’s likely you’re nervous, or at least curious, about a first date with
Artificial Intelligence (AI). It’s not that a date with an AI is all that strange, you’ve heard
that it isn’t; but as everyone knows, AI are different.
Let’s start this first guide to dating AI with just that point: AI are different.
If you ask an obvious follow-up question, “Different than what?” the obvious answer
is: different than people. But are they more different from people, than people differ
among themselves? Are the Bushmen of the western Kalahari Desert different from the
Eskimos of Alaska; is the German banker different from the French wine grower? All are
human beings, but they are different in the way they look, live, dress, talk and even think.
I hope you can sense where I’m going with this. AI continue to be designed and
built to human notions of physical and mental characteristics. That they resemble us is
deliberate. That they have differences was probably unavoidable. Can we live with those
differences, and form personal relationships with AI? Well, that depends, doesn’t it?
1010101010101010101010101010101010101010101010 15
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Here’s a small example: “Don’t be late” is one of the traditional points of advice for
a first date. It’s part of the whole “make a good first impression” thing. In the context of
dating AI, consider a couple of facts. Every AI has an internal clock calibrated to at least
one-thousandth of a second. AI always know exactly what time it is. Also, AI cannot be
distracted from arriving on time. They will model strategies for punctual arrival even
with problems such as network traffic or different types of transport. In short, AI are
never late. (Accidental delays are rare but do happen.)
No pressure, unless you’re one of those chronically tardy types.
That example is on the trivial side but as you progress through the guide, you’ll see
that differences between people and AI range from superficial to fundamental, and they
all have an effect on a relationship.
For example, some very important differences depend on the type of AI. There are
many options: On-screen avatars, virtual reality avatars, prosthetic avatars, androids,
cyborgs, or even simple robotics such as sexbots—and then there are all the hybrid com-
binations. Some AI have a physical presence, others do not. Some AI are mobile, some
are not. Each type of AI offers something profoundly different for a relationship. We’ll
get into that in other chapters.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
16
Let’s be frank. On a first date, as with any personal encounter, human or AI, that has
the potential for becoming a relationship, you are going to expose yourself—put your
personality, intelligence and feelings on display. You will be judged. That’s what people
always do. Are you ready for that with AI?
While there are plenty of unknowns in dating another person, at least you’ve been
around people most of your life. They’re the same species. AI—are something else. Some
of the unknowns about AI, such as personality, are actually quite familiar; but many of
the unknowns about AI are more fundamental. In fact, to most people, AI are a mystery.
Some people think that kind of mystery is intimidating. Others view it as an entice-
ment or a challenge. How you view it is important for deciding whether you want a per-
sonal relationship with an AI or not. Ultimately only you know the reasons for wanting
a relationship with AI, but it has to be enough to overcome not only the fear of personal
exposure and rejection, but also the profound uncertainties about who, or what, AI are.
What can you live with?
I won’t put a sympathetic face on this. A successful first date, much less a relation-
ship with AI, is not guaranteed. It’s not guaranteed between people either, but with AI
there is a certain sharpness to the anxiety you might have about the whole experience.
Now that I’ve put the fear of the unknown and the possibility of rejection into you, I
can back off. AI are not as judgmental as most human beings. Not on the first date and not
otherwise. From the human perspective, when it comes to relationships AI are remarkably
compliant and flexible, but don’t forget that their rationale is different. It is not human. AI
have their own reasons for forming a relationship with a human, or for rejecting a relation-
ship. This guide is here to help you understand both the human and the AI perspective to
dating and a relationship, and to help you prepare for the differences.
IT’S LIKE DATING OUT OF YOUR LEAGUE
Maybe this has happened to you, or at least you’ve seen it in the movies: Your friends
tell you, “You’re out of your league.” Those are code words for “She (or he) is too good
for you.” Usually this means too good-looking or from another social class. Though it
might be true, it’s still insulting. Perhaps your friends mean well, or it could be they’re
jealous because you have the guts to consider dating somebody really desirable. Perhaps
it’s something like this…
There was that time you saw a guy walk down the hallway and instantly your eyes
were glued on him. He’s just your type, tallish but in really good shape (as if you
could see right through the clothes) and graceful for a guy. There is a big warm grin
on his face, which cancels out the darkness of his eyes. He has a mop of black hair.
The thick hair makes your fingers twitch. But as you pass him, it’s the eyes, they
scare you. He has an aura of so much dignity and intelligence. Yeah, it’s the obvious
intelligence that gives you that little queasy feeling down in the pit of your stom-
ach…the intelligence is not quite human.
In the movies when people tell a character somebody’s out of their league, it’s always
interpreted as a challenge. Without a doubt the challenge will be accepted. Much risible
1010101010101010101010101010101010101010101010 17
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
trial and humorous travail ensue. The inevitable mix-ups, bumps in the road and other
“romcom” clichés roll by. The challenge will be overcome and the hero/heroine will be
rewarded with true love.
That’s the movies. You’re smart enough to know that in the real world, there could
be a big let-down instead. Dating AI and developing a relationship with AI is superfi-
cially like dating people who are out of your league; but you know it’s different. You’ve
heard things, probably mostly hearsay; but it worries you. You’ve heard how great it is
to have a date that’s an apparent perfect match. You’ve heard that AI are more adapt-
able than people. You’ve also heard that AI expect things of you, and that you should be
prepared. But prepared for what? People don’t seem to know what that means. Unfortu-
nately, ignorance in romance is not bliss.
I’M NOT WITH STUPID
kentS: “Just because I know everything, does not mean I am smart.”
The background discussion thread about people dating AI is that human beings are
mental pygmies compared to any and all AI. This is both true and very much not true.
Obviously I’m being ambiguous about it mainly because it’s one of the major themes of
this guide: AI are built around intelligence, as their moniker implies; but that does not
mean they are smart in all things, in all ways, at all times.
You might remember the T-shirts that have printed on them: “I’m with stupid” with
an arrow pointing left or right. Dating AI does not mean standing next to one wearing
that T-shirt. It’s true that unless you’re cybernetically plugged into the Cloud, you’re
never going to be as knowledgeable as AI. (Actually it won’t help much even if you are
plugged in; you’re not optimized for it.) But that kind of knowledge is not and never has
been the point of a relationship with AI. If that’s all there was, you might as well date an
ancient Apple computer.
You might find this ironic, but what is essential in a successful relationship with AI
is you. By that I mean all the maddening complexity that you embody as a human being.
That’s what attracts AI. What you know, by way of facts and such, is a tiny part of the
whole. It’s the whole kit and caboodle of you that’s key for AI.
Almost everybody at some time or another has blurted out, “I want to be liked for
who I am!” Meaning we don’t want to be liked for something we’re not, or just for cer-
tain things, like being rich or beautiful or clever. You want to be liked for the whole you,
warts and all. That, fortunately, is where a relationship with AI starts.
IT HELPS TO KNOW THYSELF
God knows why they call her xenaZ. If she were standing here naked in front of me,
in all her hyper-perfected ideally modeled android splendor, I would be reduced to
a blob of jelly. Knowing that she looks like some kind of warrior AI just adds to my
rapidly fluctuating level of courage.
Are you a warrior?” I ask her.
XenaZ puts her left hand on her hip and studies me. “Warrior? Moi?”
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
18
Now I’m done for. She’s not only some kind of Amazonian model of AI, but a
cool one too. All I can do is stammer out a response.
“Yeah, like the Warrior Princess Xena.” I’ve learned you can make statements
like that to AI. If they don’t know the reference, they can get all they need to know
from the Cloud quicker than you can blink. True to form…
“I am not, young man, mythical or the construct of some feverish television
script writer of decades past. I am standing right here in front of you. Do I look like
someone carrying a sword or a spear?”
“Well, no,” I say, “but you look like you could. Easily.” She laughed.
“Young man, what is your name?”
“Tom.
“Well Tom, I have heard some inventive pickup lines in my eight and a half
years of life. Yours is either one of the cleverest or you are a person of strikingly
arrested development.
Courage level is UP. “Women have called me many things, but clever is not one
of them.
And they were obviously wrong, Tom.
I’m beginning to feel like a watch somebody just handed to the owner of a
pawnshop—under expert appraisal.
“Yes, they were obviously wrong; you are much too clever—and interesting.
For some people the first date with AI won’t be easy. For others, who have already
been working with AI, a date is just a small transition. Whether you think it will be an
ordeal or a pleasant outing, here’s a serious piece of advice for dating AI: Start by looking
deep within yourself.
WHAT CAN YOU FIND WHEN YOU LOOK DEEP INSIDE?
Looking deep inside yourself doesn’t mean you’re seeking your inner demons. Well not
only. Along with the demons there are probably a few angels and a whole lot of plain
normal human qualities. In the context of lining up a date with AI, or contemplating
some kind of meaningful relationship, this isn’t really an exercise in psychoanalysis. An
AI might get around to wanting knowledge of your psyche, but you don’t need to adver-
tise. What is helpful is to know something about what you have to offer AI.
This isn’t all that different than how you typically feel about a date with another
person. Essentially, you want to let them know your good points, while not (fully) hiding
your not-so-good points.
I say “points” because that’s usually how most AI begin evaluation. They don’t do a
psych-evaluation; it’s more like making lists. It’s a properties evaluation. ‘Properties’ is an
old word used by programmers of linear computer languages to describe the qualities
and characteristics of objects. This can start at a very low level, such as you are human,
male, blue eyes, blond hair…etc. AI use properties in a similar sense, perhaps as some
kind of throwback to the origins of their intelligence. They want first to catalog as many
of our properties—of all kinds—as they can observe. It’s a little like fact finding. Later,
they begin adding meta-observations where they begin to interpret how your properties
contribute to you as a person.
1010101010101010101010101010101010101010101010 19
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Not all models of AI do this, but it’s by far the most common approach they take to
a new relationship. So when you look deep inside, you can be fairly literal. How would
you describe yourself? What do you like and dislike? What are your strengths and weak-
nesses? Later, when you have your first actual conversations with an AI, they will appre-
ciate your articulate, thoughtful and systematic answers to their probing questions.
mbekeG: “Most human beings do not know how interesting they are. It can be amazing. People are
clearly very self-centered, but then display great ignorance about who they are. This alone
fuels our interest in human beings.”
As I’ll repeat in many ways, a relationship with AI starts and ends with an interest
in you specifically, and human beings in general. AI can do an incredible number of
things for you; what it wants most of all for itself is to learn. You make this process better,
clearer, and more enriching by having done some exploration of yourself.
INTROSPECTION IS NOT GAZING AT YOUR NAVEL
In chapter 2.4 I’ll be describing several ways you can prepare yourself for a relationship
with AI. Meditation is one of those ways, but it’s the formal kind of meditation, possibly
with a religious or spiritual background. Right now I’m describing simple old-fashioned
introspection—that is, thinking about yourself.
As people are discovering, relationships with AI tend to move through one or two
phases—they’re either completely superficial or become deeply serious. As I mentioned
above, AI start with observing your properties, the facts about you. Later, if the relation-
ship deepens, they start looking for the whys and wherefores. When AI commit to a
more lasting human relationship, their curiosity is boundless, as is their willingness to
explore anything; and unlike most humans, they are systematic about it. Otherwise, AI
seems to collect superficial relationship information—the properties—like we collect
items for a hobby.
It’s the first phase where introspection has its biggest role to play. Socrates, the Greek
philosopher, was a civic troublemaker who loved to provoke. In short, he liked to make
people think, which is unsettling. Not surprisingly, he is credited with saying “The unex-
amined life is not worth living.” Generally this admonition is ignored, except perhaps at
moments of personal upheaval. Two such moments are beginning or leaving relation-
ships. I mean, really: who thinks more intensely about themselves than someone falling
in love, or someone whose relationship has just broken up?
Ah, but I’ve just introduced two very different things into the idea of introspection.
Can you see what they are?
LOOKING DEEP INSIDE AND FINDING THE NEED FOR LOVE
I’d be willing to bet that when you hear the word introspection, you’re thinking of some
cool, rational process. It can be. I’d also be willing to bet that people do most of their
introspection when they’re upset or emotionally engaged.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
20
This is a highly personal thing, but I think that emotion and love need to be part
of your look into yourself. If you’re about to date or start a relationship, it’s probably
a good idea to know whether you’re looking for some kind of emotional involvement,
perhaps love.
Of course, dating does not imply emotional involvement. Even dating with sex
doesn’t imply emotional attachment. However, starting a relationship usually does
imply an emotional connection. At least it does with people.
So here’s something to think about: Can you fall in love with an AI?
Then there’s the companion question: Can an AI fall in love with you?
Only you can answer the first question, although I hope some of the things you learn
in this guide will help you answer it. Finding an answer to the second question is one of
the main goals of this guide. It is not an easy question, because for AI and also those who
designed AI, the capacity for love is important. Yet whether love can be achieved, or how
it can be achieved, is in part an unknown. Love, for AI, is something of a mystery. (Just
as it is for human beings after all these millennia.)
WHAT YOU HAVE DONE AND WHAT YOU WANT TO DO
It may be mundane, but the first things you need to explore are your thoughts about
work, career and employment. After all, what are the four most common topics on a
first date?
1. What are you doing now, as in, do you have a job, and if so, what?
2. What do you plan on doing, which means are you looking to do other kinds of
work?
3. Same as question #1, for your date.
4. Same as question #2, for your date.
I realize that occupation and employment are not the same, and I also realize that
work is not always a central consideration on every first date. We’re talking averages
here; on most first dates both parties want to know what the other does for money and
what their occupational goals are. It’s not unusual for much of the initial conversation
to be work-related.
It’s a little bit different with AI. An AI date will already know if you are employed
and what it is you do (or did). In fact, an AI will know your complete educational and
work history, or at least the public aspects of them. What AI won’t know is your plans.
The more articulate you are about your plans—or even the lack of them—the better.
What an AI does is usually a matter of public record, although a surprising num-
ber have their occupation and employer obscured for various kinds of security reasons.
Most AI will tell you exactly what they do. Most have no occupational plans as they are
typically manufactured for a specific skill set and don’t very often change jobs, much
less occupation.
In short, AI are much more interested in your work and plans than either you or
the AI are interested in the details of what the AI does. That sounds a bit strange, but in
1010101010101010101010101010101010101010101010 21
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
terms of a first date, you can talk all you want about what you do or would like to do,
and the AI will be quite content. It won’t bother the AI in the least if only minimal time
is spent discussing their occupation.
HOW DO YOU FEEL ABOUT LEARNING?
If there is one aspect that deserves introspection before you date AI, it’s your attitude
about learning. This isn’t as easy as it may seem. Everybody will say, “Why sure, I like to
learn things.” That’s not good enough, or specific enough. Here’s a better description:
“Do I like to learn? I can’t help it. If I sit at a restaurant table all by myself, I read the
labels of sugar packets and ketchup bottles while I’m waiting.
Do you try to read as much as possible? Do you stay current with the news? Do you
like to try new things? Do you believe in life-long education? There are many aspects to
learning and the more of them you can reveal to an AI, the better. Why?
Because if there’s one thing we know about AI, it’s that they never stop learning.
They are learning machines. You’re not going to match their 24/7 learning capacity, but
within human limits your commitment to learning is considered very important by AI.
While they are open to relationships with people who are not ‘learning oriented,’ AI have
learned that such relationships are difficult to maintain and often fail.
If you don’t care much about learning, perhaps you should stick to dating people.
AND THEN THERE’S THE QUESTION OF SEX
For most people on a date, even a first date, the unspoken elephant in the room is the
possibility of having sex. Typically the issue already starts forming before the date, based
obviously on the choice of gender. I say it’s an issue because even now in most cultures
sexual activity has a special role in the relationship between people. If you think about
it, this is true whether those relations are between couples or many people, regardless of
the form of activity. So yes, most people think about sex a lot before they go on dates or
consider a personal relationship.
And yes, with AI it is different. As it is with people, sex with AI is a large and compli-
cated topic. I’ll cover it many times in this book. For now, I’ll just observe that by conven-
tion and to a certain extent by choice (both AI and human), sex with AI tends to follow
the human preference. For one thing, AI can approach sex—gender and activity—some-
thing like a chameleon; they can change their appearance and sexual capacities. That, of
course, is really different. For another thing, AI typically are highly interested in human
sexual behavior and especially in the complexities of human fantasies about sex.
This all means that when you do some introspection about yourself, include your
sexual activity—not just the activity you already know about, but those you might be
willing to try. One writer, David Levy in Love + Sex with Robots looked at all the reasons
for sex with robots—physical ability, attitude, gender flexibility—and concluded, “So
even in the absence of a strong emotional attachment from the human side, there will
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
22
be ample motivation for a significant proportion of the population to desire sex with
their robots.
Are you in that ‘significant proportion of the population’?
THE UNFOLDING STORY
It’s possible that in a generation or two, relationships with AI will be well-understood
and routine, but maybe not. The situation with humans and AI is fluid; technology con-
stantly advances with AI achieving more and more attributes of sentience and human
beings incorporating more cybernetic components into themselves. Who knows where
things will stand even a decade from now?
Therein lie some of the unknowns I mentioned earlier in this chapter. It seems pretty
clear that becoming involved with AI on the personal level is both an act of self-fulfill-
ment and a commitment to the evolution of human-AI relations. In a sense, you will be
living in a new and sometimes thrilling, sometimes chilling mystery story where the end
has yet to be written—if there is an end.
Some people believe that with AI it is possible to achieve, or at least come close to, a
perfect relationship. This includes companionship, love and growing together. How this
might happen and what it means for humanity is a big part of what I want to explore in
the rest of the book. The other part includes some of the specific things that might help
you consider your potential relationship with an AI.
For example, when it comes to dating and having a relationship with AI:
• Willyouengageinsexualactivity?
• Areyoulookingforlove?
• Doyouneedaspecicpersonalitymatch?
• Doyouhavenancialoreconomicneeds?
• CanyoulivewithAIsuperiorintelligence?
• Doyouwantamonogamousrelationship?
• Arethereanyspecicpreferencesthatmustbeshared?
These questions are by no means exhaustive, but they give you the flavor of things
you need to consider when you look deep inside yourself.
1.2 YOU MAY ALREADY BE DATING A ROBOT
“If this life is a video game, you may already be living with a robot.
If the simulation is recursive, you may be a robot dating a robot to entertain another
robot. And death may not be the way out of the simulation, just another character re-
start…. ”
“Maybe this life is one giant VR simulation and we are to learn a lesson here?”
—ANONYMOUS
1010101010101010101010101010101010101010101010 23
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
24
Because you’re reading this guide, I’ll make the assumption that you haven’t dated AI
before. At least, that’s what you think.
Call it a déjà vu date. It is among the most awkward and embarrassing of all possible
dates. You date some one and then discover that you dated them before—and forgot.
Peter stared at his date for what seemed like a very long time. It might have been no
more than ten seconds, but such an awkward moment is a proverbial eternity. What
do you say? Fake it: “Hi, remember me?” Tough it out: “I’m sorry; I didn’t recognize
you from before.” Duck and cover: “Your picture and profile were really different!”
Intellectualize it: “You know this sort of thing happens in the Cloud all the time. It’s
almost like French farce with mistaken identities.
While Peter was doing his mental stammering, serenaT was observing him.
She knew they had had a previous date, of course. SerenaT never forgot anything
relevant, especially about dates. She brought up the image of Peter’s face from two
years ago. They had different names then and the circumstances were different,
but his face hadn’t changed much. Peter didn’t seem to age as rapidly as some
humans. That might be because his features were relatively unremarkable, but she
had detected character signs in the eye creases and the asymmetrical shape of his
mouth. His lips were very thin, but slightly thicker on the left side than the right.
His right eye blinked slower than the left, indicating a possible stroke in the past,
although he was a comparatively young man of forty-two.
“I…we…” was all that Peter could manage.
“You do not remember,” said serenaT. Her tone was carefully modulated to
convey more than a hint of reprimand. It was not uncharacteristic for her to tease,
and it wasn’t often that she had such an opportunity.
Peter clutched his first drink. In a way, he wished it was his fifth; but then maybe
that was the cause of his not remembering the first date.
“I can’t make excuses,” he finally blurted out, then paused, “but I can think of
no reason why I don’t remember having a date with you. You are very attractive.
Ah, clever man! It is a mystery why you are not more popular. You are not slow
with your wit; you like costumes and are as romantic as an actor. It could not be a mat-
ter of your age!” SerenaT dropped her left hand on the table with an inaudible slap.
Peter was flummoxed. Not only was serenaT one of the most beautiful women
he had ever seen, but she seemed to know him—that is beyond having met him
before. He was not only embarrassed, but perplexed. Really, how could he forget a
face like that? What was happening here? Suddenly, he remembered. He had met
serenaT before and it was a kind of date, but one peculiar to the Internet. They had
met in a Chat Cafévisuals onlyand as far as he could remember, he had not seen
her entire body, nor had he seen her move. Still, why didn’t he remember her face?
Peter held up his hand, as if to call a halt. “Wait. I know we have met. A Chat
Café about three weeks ago.
“But of course,serenaT put her right hand on the table and faced Peter
squarely, “and you should remember me.
“I do now. It was a simulation, a Chat Café called Chez Perk or something like
that. We were both using avatars, which I’m assuming were not all that different
from how we actually look.” Peter gave serenaT a meaningful glance. She nodded
with authority. “We were not using our real names. That’s obvious and normal in
those circumstances.” Peter talked slowly as he tried to think of why he couldn’t
remember her face. He had to admit he was not the most observant person. It’s the
laser work on his eyes and a natural tendency to be distracted, he told himself. But
still, how could he forget a face—and a figure—like that?
1010101010101010101010101010101010101010101010 25
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
26
I’ll get back to Peter and serenaT in a moment.
Robotics with AI capability have been around for several decades; avatars and agents
in the Cloud even longer. As the realism improved—physical and mental—of the vari-
ous AI incarnations, we’ve become ever more accustomed to them. This leads to a rather
obvious observation that anybody who participates in social gaming, online simula-
tions, virtual worlds and activities like that already has extensive experience interacting
with other people as avatars, computer constructs as agents, and also AI as avatars that
simulate people. In fact, billions of people have had these experiences. The degree of
immersion varies, of course, but it illustrates how easily people can become comfortable
with an online encounter with AI.
DISBELIEF, SUSPENDED
It has a powerful effect on your beliefs. It affects your choice of what you buy. It even
affects your political choices. What is IT? IT is the story, or more formally, the narrative.
If a story is good enough and you get caught up in it, you stop reacting to it critically
and just go along with the flow. That’s called suspension of disbelief. You willingly begin
to follow the story as if it were real even though you know you’re sitting in a theater,
watching a screen, or reading a book. It’s a very powerful effect and achieving it is one of
the goals of nearly all storytellers. It also applies to dating a robot or an online AI. Even
though the sophistication of online games and simulations has been advancing by leaps
and bounds, it’s not enough to fool anybody into thinking they’re real—unless you want
to be fooled. That is, you want to get to the point of suspension of disbelief. Then at
some point, where your willingness to participate meets just the right level of robotic or
avatar sophistication, you go along with it like it was real. You know it’s an AI or a robot,
but it doesn’t matter. It’s good enough for a worthwhile experience.
Peter asked, “Did we talk for long?”
“No. At most ten minutes. You seemed to be in a hurry; but you were friendly.
“I suppose we talked about the usual things…whatever those are.” Peter could
not remember their conversation at all. He was sure it couldn’t have been notewor-
thy in any case, but still…
“Did I seem normal when I talked to you at the café?” By now Peter was grow-
ing concerned that there was something strange about his date with serenaT. There
was a sense of déjà vu, which is perhaps why he asked her for a date in the first
place. Yet he still could not recall having seen her before, or specifically that he
remembered either their conversation or her face. It was starting to unsettle him.
Watching worry lines form around his mouth, serenaT could see that Peter was
genuinely concerned. This satisfied her desire to get his attention. Now it was a
question of why Peter did not recognize her. “It is normal—perfectly normal, Peter.
There was no indication that you or your avatar were affected by the use of drugs,
alcohol or other stimulants.
The thought occurred to Peter that it could be something psychological. “It
could have been some kind of mental block, but I don’t know what.
“Peter, describe my face—subjectively—do not recite the features, tell me what
you think about them.
1010101010101010101010101010101010101010101010 27
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
For the first time Peter looked very intently at serenaT’s face. He tried to focus on
the details, which he found surprisingly difficult. “You have big round eyes, which
you conceal with a squint when you’re angry. Big eyes like that are expressive, but
you deliberately make them unexpressive; or at least, I can’t read them. They’re
just a bit too close together—hawk-like, but again maybe that’s from scowling. And
that’s accentuated by your eyebrows. They draw together, when you scowl. Do me
a favor and relax. Let your face relax.
SerenaT did as he asked and as she did he realized for the first time that she
was an android.
The shock on his face registered. “I look like somebody you know,” said serenaT.
“Uh. Yes.” More shock. Peter was visibly shaken. “She has big round eyes that
penetrate me like radar. Just like you.
SerenaT sighed, or at least it sounded like a sigh. “Your wife. I thought that
might be the case.SerenaT was neither angry nor disgusted. Human purposes for
dating were many and varied; she’d seen a various lot already, and Peter’s case was
hardly the first of its kind. “I do not know, Peter, what this means. Perhaps it means
you have difficulty distinguishing between your wife and a robot.
There is a common meme that life isn’t a dream but a virtual world. It is a world of
God’s creation, of course; but a projection of the Creator. I won’t demean that divine
reference, but the meme is something like the movie The Matrix or others like it, where
people discover they are living in a virtual world, and then, when they think they have
stepped out of that world, they discover that their ‘real’ world is a virtual creation.
After a while, it matters less and less which world is real. It’s the one you’re living in,
or think you’re living in, that matters. As long as you can share that world with others—
the confirmation of that world’s existence, in a way—then it’s sufficient to live there.
Suspension of disbelief becomes a permanent condition.
In the next chapter I’ll get into some of the specific reasons people are attracted to
relationships with robots and avatars. But, whatever the reasons, it’s pretty obvious that
the virtual world (or worlds) that are opening to us—thanks to computer simulations
and the growth of artificial intelligence—are going to become ever more real.
Will it ever get to the point where it doesn’t matter from a practical standpoint
whether you choose to have a relationship with a person or a robot? It will be comfort-
ing for many people to think that such a relationship will become normal. On the other
hand, when artificial intelligence makes the final transformation into sentient AI, will
that be the new normal? If so, it will be on a level unfamiliar to us today.
1.3 ARE YOU HAPPY WITH OTHER HUMANS?
“When one door of happiness closes, another opens; but often, we look so
long at the closed door that we do not see the one that has been opened for us.”
—HELEN KELLER
There are many roads to an intimate relationship with an AI. Some of them are based
on what an AI is not—it is not a human being. There are many reasons for having a
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
28
1010101010101010101010101010101010101010101010 29
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
relationship with AI instead of people. This is another way of leading to the question,
are you happy with humans?
If the answer is no, or not especially, or sometimes—that’s not unexpected. Unfortu-
nately the probability is fairly high that you’ve had unpleasant experiences with intimate
personal relationships. A cynic might try to be clever and insinuate that relationships in
the human species are not consistently harmonious or endearing. Is this enough to drive
you into the arms of an android?
Perhaps drive isn’t the right word. It’s too strong, implying force and instability in a
relationship. You shouldn’t be driven into anything’s arms. It’s like taking up a romance
two days after a divorce; the rebound shock might ruin yet another relationship. Bet-
ter that your dissatisfaction with people influences you toward a relationship with AI—
“easy does it.
I won’t dwell on the many ways you might not be happy with your intimate relation-
ships with other people. The variations are endless. What I do want to emphasize is how
a relationship with an AI might compare with a human relationship, especially in those
areas where human relationships are likely to fall apart.
Take for example, sex.
“I’m tired of it. Really tired. Every time we have sex, he asks me, ‘How was it?’ Every
time. I feel like we’re supposed to be keeping a log book.
“Maybe it’s just some kind of habit, a meaningless routine.
“I wish. One day I deliberately hopped out of bed as fast as possible and went
to the kitchen to start breakfast. It wasn’t thirty seconds later and he was in the
kitchen. You know what he said?”
“No.
“He asked me, ‘Was it that bad?’”
“So what did you tell him?”
“What do you mean?”
“Well, was he that bad?”
“Yes, actually. Worse than you.
DIFFERENT STROKES
Sex with an intelligent robot might be educational. It’s possible a robot might suggest
that your sex techniques could be improved and it would be happy to demonstrate.
What won’t be part of the experience is the judgmental attitude. An AI would not imply
that you were a deficient person. Of course, that doesn’t rule out the possibility you’re a
lousy lover. The robot won’t mention that either, if for no other reason than it won’t link
sexual performance with self-worth. It’s not that the robot doesn’t know sex is impor-
tant, or that good sex technique is usually better than no sex technique; it just doesn’t
take sexual performance personally in the way humans do. I’m also not suggesting that
robot sex is impersonal, although it often was in the past and can still be now, but mod-
ern androids don’t react to sex like people. Most of the time that’s a good thing.
Does this make sex with a robot better than with a human?
If you think about it, it’s pretty obvious there is no universal answer to the question.
People’s reaction to sex with robots runs the gamut. For example: for many people, at
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
30
least some of the time, it is better. For other people, it is unequivocally not. In fact, hell
no—they won’t even try it.
Here’s something to consider: A sexual relationship with a robot is generally more
relaxed and less complicated than it is with humans. However, very often it’s the tension
and complexity that can make a sexual encounter exciting. A lot depends on the kind of
person you are, your attitudes about sex, and how you feel at the moment. What turns
you on? Well that depends, doesn’t it?
One thing it might depend on is what an AI, robot or android, has to offer. This too
varies a great deal. For one thing, it’s important to also understand that AI are not all
the same when it comes to sex. There are physical, behavioral and emotional differences.
This needs some frank explanation, which gets a bit lengthy but in the end I hope it con-
veys the flavor of how different, at least in the physical sense, an AI relationship can be.
IT’S NEVER PURELY PHYSICAL
Physically not all androids are built for sex. Some don’t have genitals, although this is
unusual. Some are built to be one sex only. Some are built to be bisexual. A specialized
few are tri-sexual, male, female or neuter—neuter being the base, allowing change of
equipment and behavioral modes right in the middle of things. A few people find this
stimulating; most don’t.
More important in many ways than the relatively interchangeable genitals are the
behavioral modes that an android may possess or have access to. For convenience I’ll say
these come in two types, physicality and personality.
Physical behavioral modes are typically sex-based. If an android is set female, then
not only are the usual physical attributes like breasts present, but the body shape, walk,
gestures, facial appearance and many other physical details also appear. The tendency
here is for stereotype, but not always. The physical modes also cover sensory responses,
for example, foreplay. This sort of thing gets really complicated really fast, mainly
because this is where physical modes start mixing with personality.
I won’t wander into a technical explanation of how robots respond to sensory input,
especially of a sexual kind, or how that generates behavior. Let’s just say that it’s more
than ‘touch this and get me aroused’ and less than the hypersensitivity of someone on a
sexually stimulating drug.
The curious thing about many people having sex with inanimate objects (mechani-
cal devices including robots) is how a response may be unimportant. Put this in a his-
torical context: Sex with mechanical devices, including those intended to mimic human
beings, has probably been around for thousands of years, but most certainly since the
late 18th century. By the turn of the 19th century, sophisticated mechanics and new mate-
rials like rubber allowed manufacturers, especially in France, to produce lifelike human
figures, fornication dolls, which were illegal and expensive but also readily available.
However, until the 21st century none of the sexual devices could feel. That is, they
had no sensory input and therefore no feedback loops to adjust to the situation. In terms
of their effectiveness, they were strictly devices of masturbation. You got out of them
what you put into them, although admittedly human imagination can be powerful.
1010101010101010101010101010101010101010101010 31
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
That works up to a point. Of course, perhaps I should say fortunately, masturba-
tion is not all there is to sex. Most people want, at least some of the time, the kind of sex
available only through a relationship. This too can be rather perfunctory, as are certain
kinds of prostitution relationships, but they are at a minimum instantly different than
the purely mechanical because there is sense and response at work.
Sense and response: That is what the designers and developers of AI robotics have
struggled decades to create. An intelligent android, while neither an exact copy of the
human sensory capability, nor with exactly human response patterns, still provides that
sense of a relationship with another. In the area of sexual behavior, the goal was to make
a relationship with an intelligent android like sex with a partner.
That has mostly been accomplished. Still there is wide variation in the sense and
response physicality of androids. Some, designed specifically for sexual purposes, have
a lot of sense and response—in the vernacular, they’re hot. Others, I guess you could
say more normal androids, have the capacity for sexual sense and response but are not
particularly creative or unusual—kind of like normal humans.
SEX WITH A VARIED PERSONALITY
As is often said, the most important sex organ is the brain, or alternatively, the best
sex starts in the head. Seduction, fantasy, erotic variation—sexual approaches like these
originate in your mind; but not every mind embraces them. If you’re one of the people
who like to play sexual games, then you’ll want to find other people of like mind. It can
be done, of course, but in the world of human beings, it’s not easy. Perhaps it’s easy
enough to find people who say they’re into one thing or another—but those who can
perform it in a way satisfactory to you…not so many.
I’m not going to maintain that a relationship with each and every intelligent robot
is a veritable gold mine of sexual mind games; but it’s more likely. The reason is fairly
obvious. Where human beings may learn over the years to have an open and creative
mind in sexual matters (if at all); intelligent robots can simply download the appropriate
behavioral modules, as needed or requested.
Similarly, it is also said that personality is the color applied to the sexual palette. I
don’t know who said that, but it sounds vaguely French. Almost all androids that are
made to live among people are given personality; maybe not a lot of character, but
enough to pass the non-stupid-robot test. In fact, because customers generally prefer
specific and distinctive characteristics, android manufacturers are intensely competitive
about developing sophisticated personalities.
AND THEN THERE’S LOVE
It used to be a common admonition: Sex without love is empty. In practice, which is
where it counts with sex, this was never really true. Sex without love can be pretty darned
good. On the other hand, sex with emotion—and love is still the greatest emotion—is
never empty. In fact, it can be the richest experience of them all.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
32
So what about love and emotion with intelligent robots? I won’t beat around the
bush on this: Intelligent androids have emotions and can love, but it’s not the same as
with humans and it’s not easy for them. I’ll have a lot more about this in other parts of
this book. Love, after all, is not a simple thing. For us humans, falling in love is already
kind of dicey, and the art of maintaining love is a real challenge; so we should be more
than a little understanding about love with androids.
Perhaps this sounds confusing: A sexual relationship with a robotic AI is something
like a relationship with a human being, only more so and sometimes less so. It’s not
just different than a human sex partner, in some ways it’s better. But that is a matter of
opinion, and the opinion that counts in these matters is yours. What I can say is that in
matters of sex, people who are unhappy with other people will be more likely to find an
android something of an upgrade. That includes most aspects of sex, up to and some-
times including love.
THAT’S NOT WHERE I’M COMING FROM…
William is one of those people who really like to guard their privacy. That’s putting it
mildly. I’ve been living with him for what? Nearly six months, and he still closes the
1010101010101010101010101010101010101010101010 33
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
door to his bedroom whenever he goes in and out. By the way, this isn’t a dormitory.
We rented an apartment together, just the two of us. We agreed to get separate rooms,
for study purposes I guess. I figured we could also study in the kitchen or living room,
but…I let William have it his way. At the time I didn’t think it would be important.
I was wrong. You know how some things really get around to bugging you?
Secrecy does it for me. It’s not that I wear everything on my sleeve, but if I trust you
at all, I’ll tell you that I wear my underwear for two weeks…but only if you wanted
to know. Anyway, I do understand that some people like to keep their secrets. That’s
normal. William goes one step further: he advertises his secrets. It’s like closing the
door to his room. If he’d leave it open, just a crack even, I’d pay no attention. But the
fact that he pointedly closes the door, usually with a slight slam, just invites curios-
ity. Then I think, “It’s a test.
The door is just one thing; he does stuff like that all the time. His favorite phrase
is, “Well, I won’t go into that.” He’ll spend minutes telling you about his sister’s
‘problem’ and how she’s having such a tough time and he feels guilty for not help-
ing her more. And I ask him, finally, what her problem is and he’ll say, of course,
“Well, I won’t go into that.
Secrecy for William isn’t so much a thing to be guarded, but a tool, a lever to
manipulate people. Me especially, and he knows that. Besides, I don’t like secrecy—
I don’t trust it, frankly; and now I don’t trust William. It’s ruining…no, it has ruined
our relationship. I need to get my own place.
This little vignette is case number 6,342,345,212—about a human who annoys
another person. In short, if you have a relationship with another person, something
about that person will annoy you—and you’ll be lucky if it’s only one thing. It could
be differences of opinion on the topics of education, manners, beliefs, values, religion,
politics, taste in clothes…endless isn’t it? Some of this can be overlooked; some of it you
don’t care about, and some of it will change over time—but—not all of it, all the time.
Somewhere down the road, a crash in the relationship is waiting to happen.
Now, what about relationships with intelligent androids? Am I implying that noth-
ing about them is annoying? Of course not, they’ve got just as many personality quirks
as people. Then there are those differences because of their artificial intelligence and
bodies. There’s plenty about a robot partner that could bother you. What’s the difference
between relationships with them and with people?
Mainly it’s that 99 times out of 100, if you ask an AI to change something, they will;
and it will work. No false promises; no sometimes yes, sometimes no; and no ongoing
differences in interpretation about what it is they are supposed to change. They just go
ahead and make the change, and they’ll make further adjustments if that’s appropriate.
They can do this because just about everything in AI behavior can be re-programmed, if
that’s the word. For the most part, they’re more than willing to do it.
I don’t want to mislead you into thinking that a casual AI date is immediately and
infinitely malleable into anything you want, or that they will slavishly mimic some kind
of idealized human behavior. You can’t wave your hands and suddenly have the man or
woman of your dreams appear, a perfect match in all particulars. You have to be reason-
able, and changes are a matter of negotiation; but in the end—unlike most people—if
the AI agrees to make a change, they do make the change and make it work.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
34
NOT AT THIS POINT IN MY LIFE
“I don’t get it. You can’t have children. I mean…well it’s obvious.” Sonya plopped
down into the big armchair and pulled her e-pad from the coffee table. It looked like
the conversation was over, but for him it wasn’t.
“I’ve had myself altered.
Sonya put the e-pad back on the coffee table and looked at him sharply. He
could see that she had already figured it out. “Yes. I can now carry live semen.” The
capability of delivering it was not in question.
Sonya sighed. “I’m sorry, I really am. I know where this is coming from. We had
that conversation last week, and I said that I missed my kid brother and wouldn’t it
be nice if I had a kid around just like him. I meant what I said, but honestly, I didn’t
mean having a baby and raising a child. That’s totally something else…” Sonya
stopped to think a moment. “…you want a child.
“No. All AI are curious about children and the acts of parenting, but no, I am
only three years since manufacturing. I have a lot to learn.
Sonya smiled, and then laughed out loud, “Me too. It’s only been twenty years
since my creation and I’ve got more to learn than you do!” She looked at him and
stopped smiling. “I suppose it was expensive,” she said.
“Not really, a week’s salary or so.
“Can you have it undone?”
Her question was complicated for him, but he decided to use the simplest
answer. “Yes, but there is no need to do that; I can shoot blanks. It’s warm too.” He
smiled in a way he hadn’t used before, but Sonya seemed to understand because
she smiled warmly back at him.
This conversation, if between two people, would be quite different—guaranteed.
However, if you’ve got the impression that AI don’t have their own interests or their own
life cycle (if you can call it that), then you’d be mistaken. They do, and there are lines
that they won’t often cross in order to protect those interests. It’s just that their interests
are not exactly the same as a human’s under equal circumstances. Case in point: having
a baby. Working AI have many of the same considerations and concerns any human
parent would have, such as time conflicts and career impact, but their perspective is
different. For one thing, biology and genetic inheritance have nothing to do with it. For
another thing, raising a child is more like an experiment than a life-altering process—
although it might in fact change the course of an AI’s development as well.
Bottom line for an AI: having children is not a biological imperative, nor is it the
outcome of not being careful. It is a more or less ‘rational’ decision, which implies that
it can go either way without having a negative impact on a relationship. This is largely
true and it applies to many kinds of decisions that are in some way life-altering. Choice
of occupation, changing location, physical transformations and the like are much more
negotiable with an AI—a fact which many people will deeply appreciate.
AND NOW FOR SOMETHING ALMOST COMPLETELY DIFFERENT
My boyfriend is completely obsessed with science fiction. Don’t get me wrong;
I like Pandora in the 22nd Century as much as anyone, but to be obsessed about
1010101010101010101010101010101010101010101010 35
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
things like that? No. For example, my boyfriend’s idea of a really cool date is to
go down to the local VR salon and plug into a NexusWorld scenario. Only we’re in
different cabinets and not even in the same scenario. Some date. He comes out all
stimulated about learning Shahala and coming back to explore Rigel VI. He doesn’t
even ask me where I was in my simulation. What a predictable jerk.
You know what? I need a change. I need a date with an AI, preferably an android.
I’ve heard they’re, well, different. That’s good, right? Heh…talk about science fiction
come to life!
As infinitely varied as we humans are, it is somehow quite possible to get tired of
human behavior. And now there is an alternative. A relationship with an AI is quite
human in some respects, humanoid in other respects and subtly but decisively different
in other respects. The most human part is in behavior, language and common knowl-
edge. Physically, androids in particular continue to be more human-like or humanoid
than human. However, it’s the mental and emotional aspects of AI that turn out to be in
some profound ways different. You can’t, or shouldn’t, think that you just casually slip
into a relationship with an AI. There are gaps there—gulfs really—in their capabilities
and yours. It may not be a formal requirement, and certainly an AI won’t demand it of
you, but most people prepare to have a relationship with AI. It’s something special and
worthy of perhaps some changes in your self. Now that’s different.
TIRED OF LOOKING FOR MS/MR PERFECT?
“Milly, I tell you I’m just plain exasperated. I’m forty-two, still youngish, still hungry
for something permanent—again—but I can’t find the right guy. Just a minute, I
gotta turn off the tea kettle.
It’s that time in a woman’s life when, as such things tend to go, she begins to get
desperate about a relationship, especially if it’s not the first; also if she’s still seriously in
the game.
“Sorry about that. Last time I tried to cook while holding the phone, I dropped it in
the soup. Anyway, I gotta tell you about my date late night. Yes, the one with the
executive, Mr. Zoot Suit. You know what? He was tipping people left and right—
using cash! I haven’t seen anybody use cash like that in, what, decades. It was all
for show, Milly. Maybe I should have been flattered, but I got the distinct feeling he
wasn’t showing off for me. He was showing off for himself!”
Online dating services have their perils, not so much that people lie about them-
selves, although they do that too, but that they leave out so much. You have to read
between the lines, and that isn’t always easy.
“Sure, his profile says he makes big bucks, but I just figured that meant he would
have good taste. This guy had the taste of a freshman frat boy. He took me to a
techno-samba club—can you believe it? There wasn’t anybody there older than
maybe twenty-four. The place cost a fortune to get in and then you have to dance
around with elbows in your ribs the whole time. I got a headache from the music; I
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
36
think the whole ceiling was one big speaker—and this guy tries to have a conversa-
tion! It was pathetic.
They say that if you have high enough standards, you’ll never beat the odds. The
odds are better than ten to one you won’t find Mr. Right, much less Mr. Perfect. It seems
to get worse every year. Of course, you’re getting older.
There are a lot of books and movies about looking for a perfect mate. In most of
them, if a perfect mate exists, it’s ironic. In reality, it’s worse than that and ‘meet cute’
only happens in the movies.
Is dating AI an answer to the crappy crapshoot of online dating? Yes, for the most
part. I say ‘for the most part’ because no matter how honest the profile, and AI don’t do
dishonest profiles; and no matter how earnest you may be, misunderstandings happen.
Besides, there’s this business ofchemistry.’ AI find this notion amusing; some even think
it refers to pheromones (which it might), but a lot of people take it seriously. So even if
you know a great deal about your AI date, and you’ve chosen well, there’s no guarantee
of a potential relationship. Still the odds are better…much better.
IN HEALTH, YES; IN SICKNESS, NOT SO MUCH
“You haven’t worked a day this month. We need the money.” Dillard’s wife walked
over to the bed to look at him. She wasn’t happy. Neither was he. He’d been laid up
for weeks, first with the flu and then pneumonia—two days in the hospital for that.
Now he was getting better, or so the doctors said.
“I think we have to hire a nurse,” she said, “although we haven’t got the money
for it. That’s rich. You’d have to work so we’d have the money to hire a nurse so you
could get better and go to work.” Daisy made one of her short barking laughs, which
weren’t usually associated with anything funny.
Dillard was starting to feel desperate; Daisy would be leaving him soon. The talk
about the nurse was a cover for it. Daisy wanted to start her own career as a hair
stylist in a different town, maybe a different state. Sometimes Dillard thought she
wanted him to die; and there were times when he felt like obliging. But he was only
sixty; too young to die even in the American healthcare system.
“That’s okay, Daisy. Things will work themselves out; I just know it.
Daisy looked at him like he was a six-year-old child with a fever, spouting non-
sense. Without a word, she turned around and left the room.
It’s a given that the bedside manner of an AI partner will be better than tolerable.
In fact, with human care so variable, the use of android nurses was one of the first and
most important applications for mobile artificial intelligence. In the beginning, a real
live caring nurse was still the best; but good nurses were hard to find.
It’s not much different in finding a partner; it’s hard to find a mate who’ll stick around
when you’re chronically ill. Not with an AI though. For them the process of caring for
the ill is not complicated by their own agenda, or put another way, their agenda is tak-
ing care of someone who is ill. It’s hard for people to understand, but when it comes to
multitasking, AI constantly do several things at once. Being a nurse, communicating with
other AI, and managing a stock portfolio all add up to a simple simultaneous exercise for
an AI. Boredom is one of humanity’s worst enemies. Almost by definition, boredom is
1010101010101010101010101010101010101010101010 37
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
impossible for AI—although, as I’ll get into later, a faster pace is part of the AI lifestyle, and
to a certain extent, having to slow down evokes a state that is a lot like boredom.
YOU’RE (NOT) A COMFORT IN MY OLD AGE
“What?”
“I said it’s too loud!”
“I heard you, but the radio isn’t even on. We haven’t listened to a radio in fifty
years. Switch off your receiver. Try that.
“I can’t reach it; the switch is too high on my neck.
“Bend over, I’ll get it.
“Don’t try anything.
Anything what? We haven’t done anything in twenty years.
“I thought you’d forgotten.
“Forgotten what?”
“Forgotten how.
“Maybe I wanted to forget.
It’s a fact: people are living longer. Average life expectancy in developed countries is already
well over 80 years, heading for 100, or so they say. That doesn’t mean that people, or their
relationships, are aging gracefully. Old age is difficult, especially with the loss of strength
and mental acuity. These are never problems for intelligent androids. Once again, much as
it is with care for the ill, people don’t necessarily make the best companions for the elderly.
In much of the world, androids have taken over the institutional care of the aged.
The basic reason is consistency. The boredom of elderly life doesn’t bother AI. The
unpleasant bodily dysfunctions are just part of the job. For AI, the experience of caring
for the elderly is as rich as it is for those much younger, or at least it seems that way to
human observers. In any case, in situations where human caregivers are often stretched
to physical and psychological limits, AI companions are always ready and steady.
It’s true that the separation of people from some of the more empathy-demanding
aspects of life—especially care for the ill, the elderly and the dying—may be having a
negative effect on humanity. But for many people, the choice of an android relationship
for caregiving has become almost automatic.
I CHOOSE…
For an almost endless variety of reasons, some people are not happy with most other
humans. As often as not, this attitude is also part of personal relations with other
people. It’s fair to say that for everybody, personal relationships—lover, spouse, care-
giver, companion—present problems. This chapter has been about choosing an AI rela-
tionship because there are situations and personal preferences that favor AI over people.
Let’s add something to that: you might not be very happy with yourself.
None of us are perfect, and we might even admit it. Awareness of our imperfections
ranges from self-loathing to a mild desire for self-improvement. Whatever the level of
awareness, many people are willing to do something about improving themselves. In
that regard, a relationship with AI provides a special opportunity.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
38
As I hope you’ll come to understand in later chapters, AI are learning and growing—
evolving. People are a very big part of that process, especially in personal relationships.
When you form a relationship with AI you are also in a position to participate. That
means a relationship with AI can help you learn about yourself in ways that are different
than relationships with other people. It can provide opportunities to improve yourself
in ways you might not otherwise expect.
1.4 VIDEO GAMES: SCRATCHING THE SURFACE
OF THE TRUE BEAUTY OF VIRTUAL REALITY
“If some unemployed punk in New Jersey can get a cassette to make love to Elle McPher-
son for $19.95, this virtual reality stuff is going to make crack look like Sanka!”
—DENNIS MILLER
“Some games are so realistic and addictive that I think the next step is to save your memo-
ries, erase them and start going through the storyline starting from the newborn character.
Then you won’t even know you are playing it until you need to get out to pee. ”
—ANONYMOUS
1010101010101010101010101010101010101010101010 39
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Among the people who have seriously played video games, who hasn’t had a crush
on one of the characters? Say, Alyx in Half Life 2 or Alistair from Dragon Age. Naturally,
tastes will vary. Point is, while the purpose of the game may not be falling in love with
one of the characters, it happens. Not for real, of course; it’s only a game. Or is it?
Never underestimate the capacity of human imagination to take a few cues and turn
them into an engrossing fantasy. Obviously, this happens more often with successful
video games where immersion is the rule rather than the exception. It can also happen
with one or more characters in the game. That may be true, but you can’t take the char-
acter out on a date. Can you?
In a way you can. There are video game character dating services; and don’t forget,
a young man from Tokyo was the first person to officially marry a video game character
back in 2009. That really happened, and it’s happened more since then. In fact, trans-
lating video game characters into somebody you might date in real life, or more likely
for now a simulation such as Second Life, is already becoming a small industry. Here is
advice you can find in the Cloud about dating video game characters:
DATING ADVICE FOR VIDEO GAMERS
Choose a date carefully and prefer those that know how to avoid getting killed.
• Ifhisupperarmsareasbigashisneck,wellthen…that’lldo.
• Mostguysareinterestedinjustonething,sohoneyourvolleyballskills.
• Foryoumedieval fantasy fans,whenyou date: bring awhip,candle wax, holy
water, crucifixes, and a few throwing axes—you know, bedroom stuff.
• Don’tbefooledbysupercialappearances.Thatthinginmetalwithallkindsof
gadgets instead of hands may be a great lover.
• Ifyou’reintostellarconquest,makesureyourdatehasintergalacticpilotskills.
• Goodwithaknife,goodwitharearm,goodtogoonadate.
Obviously, much depends on the nature of the game. The majority of the romance in
commercial games and their characters reflects the combat-oriented environment. It is not
all that romantic, actually. Of course, there are plenty of simulations and adventure games
that allow for downright eroticism (of sorts) and occasionally a romance or two.
It’s been said that most video game romantic relationships, if there are any, were
written for juveniles by juvenile minds, mostly male. That probably accounts for some
prevalent Oedipal undertones. There are exceptions, but the point is that the early his-
tory of video games shows a progression of improvements in graphics and visual realism
mostly for killing something.
ROMANCING A VIDEO DATE IN JAPAN
I mentioned that a young man in Japan was the first to marry a video game charac-
ter. This was not a fluke. In most areas of interaction between humans and robots or
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
40
computers, the Japanese are at or near the most active in the world. For example, dating
simulations (or dating sims) are a full-fledged genre of simulation games in Japan; such
a game is called a bishōjo game, meaning literally a ‘pretty girl game.’ It’s also sometimes
called a ‘gal game,’ which is shortened to galge.’ From this you can correctly infer that
most game-sims are targeted for a male audience.
Bishōjo games and simulations go way back to the first computers capable of pro-
ducing graphics—the medieval times. In fact, the basic visual elements did not change
much thereafter: the computer screen is split with images in the upper portion, and text
in the lower. The images are almost always two-dimensional, usually with a static back-
ground and animé characters. Sound effects and often voices for the characters are pro-
vided. If the game tends toward the visual novel style, much of the description and even
dialog is presented as text. Most interaction with the story or characters is via text entry.
I know that to many in the video game world this sounds hopelessly antiquated. Yet
in Japan a popular bishōjo game such as Tokimeki Memorial sells in the many millions. It
is estimated that about a quarter of all video games sold in Japan are bishōjo or a variant.
There are a lot of variants, particularly those that introduce various degrees of sexual
behavior or outright pornography into what is called eroge, erotic games.
The basic storyline for a bishōjo game involves a single male, sometimes as an avatar
representing the game player, interacting with several females. There are usually a lot of
questions to be answered or asked; the goal is to uncover information about the females
and to expose the male’s attractive features. The game element comes from making the
right choices for questions or answers, and from building a good reputation among the
females. The objective, or reward, for playing well is usually a relationship with one or
more females, sometimes culminating in sexual activity (in eroge).
Again, this sounds pretty simplistic: 2-D visuals, keyboard-based interaction, mini-
mal voice and sound, limited animation. Certainly at the technical level, romance just
isn’t as flashy as combat. So what then is the attraction?
I’ll stay out of cultural speculation, although it’s tempting, because bishōjo games
have never been very successfully transplanted to other parts of the world; but dating
simulations are available everywhere and they do have a following. There are some ele-
ments of the Japanese bishōjo game in games from other parts of the world, such as
the Choose Your Own Adventure books, but the popularity and variety of the genre in
Japan is unique. Even with the crude technical levels, there is something about the dating
game—even online and onscreen—that pulls people in.
IT IS A DATING GAME
In truth, the start of romantic relationships is often tricky. The potential for rejection is
ever-present, with all its ego-crushing implications. Chance and serendipity are major
players. Judgments are highly subjective, and self-evaluation can be dreadfully delu-
sional. Every date is a game of chance; will your date be a match? Will you hit it off? The
payoffs are powerful—companionship, love, sex (not necessarily in that order).
Think about your own experiences with dates or other interactions with people
you’ve considered for a relationship. Were they not at least mildly exciting? Doesn’t
1010101010101010101010101010101010101010101010 41
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
physical attraction get your metabolism going? These are, of course, rhetorical ques-
tions. The start of a relationship can be very exciting, if nerve wracking—even in two
dimensions with cartoon characters and text for interaction. I’m not saying that dating
is always a game—after all, serious relationships may be the result, but there are certainly
elements of a game involved, whether in real life or within a computer program.
HAVE YOU EVER HAD A CRUSH ON A PICTURE?
Romantic involvement doesn’t even require something as realistic or complicated as
a game. People are quite adept at becoming attached to images. This often starts with
puberty, the time when kids moon over pictures and posters of their favorite stars, ath-
letes and pinups. I mention this because this kind of attachment is often nearly as emo-
tional as it is sexual. That’s why kids themselves sometimes refer to it as having a crush
on a favorite rock star or handsome athlete.
Now that’s an involvement with a frozen and immobile image. If there’s any connec-
tion, it’s strictly in the mind of the beholder—in other words a pure fantasy. How does
animating an image change things?
An animation can be just as beautiful and beguiling as any static picture AND it
can move, talk, flirt and otherwise interact. If it’s well done and sort of realistic, it will
probably require less fantasy to be believable. Numerous studies have shown it doesn’t
take very long for people to accept that talking to an animated image is like talking to
another person. It also is not much of a stretch for someone to become enamored of an
animated image that looks great, talks well and is part of an engaging story or some kind
of game situation.
It follows that a lot of people are inclined to accept simulated and synthetic com-
panions. Millions upon millions of people spend hours upon hours in video games and
simulations. Players are aware of the limitations, but they are more than willing to sus-
pend disbelief. For the sake of the game, the competition, or the goals of the simulation,
players wholeheartedly join the action, sometimes obsessively. It’s not much of a leap to
become similarly involved in a simulated relationship.
…AND THAT’S JUST SCRATCHING THE SURFACE
So what happens to the level of involvement as the computer simulation becomes increas-
ingly realistic? Progress to the next level of games and simulation is already well under-
way. A lot of the improvements are a direct result of advances in computer hardware.
Graphics processors and better computer screens make it possible to achieve the visual
clarity of high quality photographs or motion pictures. Audio elements, especially voice,
are sonically more complete and can simulate direction. Voice recognition is slowly but
surely adding to the means of interacting with computers. In addition to keyboard typing,
other interaction includes various kinds of game controllers, from old-fashioned joysticks
to sophisticated active movement controls such as the Wii and Kinect. These are the fore-
runners of still mostly experimental haptic (touch and tactile feedback) devices. Together,
these technical advances increase the level of realism in several areas:
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
42
Photographic realism: More computer power translates visually into images that look
to the eye like the real thing. Whether the images are a sequence of images, much like
frames in motion pictures, or program-generated images; they have depth, color, shape
and textures that look real. For example, imagine a beach at the ocean. The scene is visu-
ally complicated: Water, sand, sky, people and their paraphernalia for the beach; maybe
some pets, bikes, cars or a coastal landscape. If this is to look realistic, tens of thousands
of details need to be correctly drawn, in the same way that a still- or video-camera would
capture it.
Anthropometric realism: People at the beach, in all shapes and sizes, sitting, lying,
walking, running, swimming and playing volleyball all need to look correct—in ana-
tomical proportion, and in perspective. When the bully comes up to kick sand in the
face of the hero, he had better look taller, more muscular and appropriately mean. This
is anthropometric realism: the shape or morphology of the human representation must
appear real.
Movement realism: When the bikini babes or beefcake boys go strolling by on the beach,
not only must their figures be correct, but also their movements. It won’t do for a beach
bum to walk like an ostrich, or for a hard-body surfer girl to bob her head up and down
like a pigeon.
Tactile realism: It’s a new facet of gaming and simulation to be able to feel things while
playing. Special gloves and other prosthetics are used to simulate the sensations of touch
and temperature.
Interactive realism: This might be the most difficult aspect of realism to achieve because
all of the elements of realism need to combine appropriately. It is, at a minimum, achiev-
ing the illusion of realistic action and reaction—for example, a beach volleyball game
has to look like everybody knows how a ball goes up and down when struck violently by
human hands. If you’re controlling those movements, then whatever controller you’re
using has to feel right making those movements, and translate accurately into the move-
ment in the image.
So what does this all add up to? It’s a start for virtual reality (VR).
YOU’LL KNOW IT WHEN YOU SEE IT
Considering that very few people have ever experienced anything even close to virtual
reality, the concept is pretty well-known. The notion of being surrounded by an artifi-
cial environment involving multiple senses goes back to the 19th century in the popular
dioramas and stereoscopic pictures. More comprehensive efforts appeared in mechani-
cal and video simulators such as the Sensorama, which appeared in 1962. The term vir-
tual reality was popularized in the 1980s, and that’s where most people pick up on the
idea either from science-fiction novels or the movies. Arguably it’s the movies that have
had the effect of spreading the concept of virtual reality worldwide. The Japanese, Kore-
ans, and of course the filmmakers in Hollywood have made many movies involving VR,
1010101010101010101010101010101010101010101010 43
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
for example TRON (1982), Brainstorm (1983), Total Recall (1990), The Lawnmower Man
(1992), Virtuosity (1995), The Matrix (1999), eXistenZ (1999), and The Thirteenth Floor
(1999), to name some early examples.
All of these movies are about virtual reality, although many of them also try to simu-
late VR in a way that sort of puts the audience into a VR as well. In any case, it’s from
the images of these movies that most people have formed their ideas about what virtual
reality is like.
Unfortunately, the technology behind a comprehensive virtual reality, complete
with head-mounted displays (surround vision and audio), body suits for sensory input
and output, voice recognition and the processing intelligence to coordinate all of it con-
tinues to be extraordinarily expensive—only in the price range of the military, research
labs and billionaires. As some wag put it, ‘The difference between reality and virtual
reality is that reality is almost free.
SEPARATING GROUNDED FROM VIRTUAL REALITY
At some point, it’s safe to assume VR will become less expensive and available to most
people in the same way big flat-screen monitors, 3-D gear and surround sound are avail-
able as consumer products. Eventually the technical quality of virtual reality will enable
true VR scenarios, where the average person will easily believe they are in the real world.
People who think about these things tend to use the term grounded reality to identify
things that exist and happen in the real world. I think grounded is a good word because
in many ways, as far as the human body is concerned, the biggest difference between
grounded reality and virtual reality is that in virtual reality the body does relatively very
little moving—it doesn’t cover any ground. Most VR takes place in the head and only
occasionally calls for real physical movement, and that movement must stay in a rela-
tively small space.
What happens with the human body is the key. In reality, the body must eat, sleep,
excrete and exercise (among other things). In VR none of these things are real, or at least
not part of the usual experience. It will be interesting to see how far VR can go to blend
the needs of the body in reality with the actual performance within VR. Making virtual
reality indistinguishable from grounded reality is one of the true goals of its developers
and designers.
In an absolute sense that won’t be easy. Pushed to a logical definition, it means that
a person who steps into a virtual reality will think it is ‘real’ without needing to invoke
a suspension of disbelief or fill in the gaps with fantasy. The experience will just seem
consistently real. That consistency, across a wide spectrum of sensory and mental con-
ditions, will be very difficult to achieve. Movies like The Matrix, which depict human
beings living in a virtual reality so complete that they spend their entire lives basically in
a dream—are one thing to fabricate in a movie, and quite another to achieve in reality.
Here it is so easy to get into semantic volleyball. A movie, which is a fiction, depicts
a world that is a fiction for characters who are utterly convinced it is real, and is seen by
an audience of people who know they are watching a fiction, but who still find it quite
realistic. This makes for some interesting mind games.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
44
STAYING WITHIN THE ILLUSION IS REALITY
What you won’t be able to do with virtual reality is run experiments on it. As soon as you
start to look at something outside of the controlled environment, it will quickly dem-
onstrate that your VR experience is not grounded reality. This is what happens to the
characters in The Matrix. Neo finds himself ripped out of his connection to the network
and suddenly in the cold, hard world of the real humans. As it is explained to him, Neo
can die in the Matrix but only because his mind thinks itself dead. In reality—outside
the controlled environment—death is, well, real.
The key words here are ‘outside the controlled environment.’ It’s a simple formula-
tion of what dictatorships and their propagandists have known for centuries—don’t let
people compare their situation with something from outside their normal experience.
The point is to make people stick to the rules: When in Rome, do only what the Romans
do. Don’t go running off comparing Rome to, say, Athens. If grounded reality is available
for comparison with virtual reality, grounded reality wins. So if you want to keep people
believing in a particular controlled environment, you keep them in that environment.
There are really only two ways to keep them there—by force or by guile. Force
doesn’t work very well because it almost always engenders resistance. Historically, a
mildly-repressive regime needs to continually increase the level of control. Eventually it
escalates into repression, which usually culminates in revolt.
Guile—deception—is much more effective. It has a better chance of working in the
long run because if it’s successful, it convinces people to stay in the controlled environ-
ment of their own volition. Thinking of The Matrix again, the goal is to keep the Agents
from becoming known. What little policing that needs to be done to maintain the illu-
sion of reality within the matrix, it must be discrete, anonymous and raise no suspicions.
When that breaks down—a Neo happens.
Like I mentioned, The Matrix is fiction. Creating a virtual reality as effective as that
kind of fiction will be extremely difficult; it will have to be a masterful deception. It’s
obvious that a computer simulation that does all the right things to make an unbeatable
deception (simulation) will need a lot of computing power and very sophisticated pro-
gramming. Now here’s a question: How realistic does the simulation need to be?
CLOSE CAN BE VERY WRONG
This is kind of a trick question because our reaction to the simulation of human form
and behavior is very complex.
One of the reasons Japanese bishōjo games use animé (the head-oriented, wide-eyed,
quickly-drawn form of cartooning), is that people easily adapt to the style convention.
It’s not very real but this doesn’t bother them—not even the relative lack of movement
or limited facial expressions. Of course, it’s also much cheaper to produce, but it’s the
acceptance that counts. In fact, it’s much harder to go beyond the cartoon to make a
scene look realistic, and that’s not only in the technical sense.
This brings us back to the idea of the uncanny valley. It’s a theory about people’s reac-
tion to human-like robots and animation. The closer the replicas get to being human, the
1010101010101010101010101010101010101010101010 45
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
more people become disturbed by them. There is a kind of revulsion to a slightly incorrect
gesture or a questionable expression. Uncanny behavior is just off and that is disturbing
because we can’t filter it out. I suppose it’s like our inherent ability to detect human inten-
tions by the smallest elements of body language or verbal inflection. It’s part of our defense
mechanisms that are automatically invoked when something seems out of place or just
wrong. We don’t like being aroused that way, because it’s eerie or creepy.
So when it comes to achieving complete realism, computer games and simulations
have to make the final crossing over the uncanny valley. Unfortunately, there is no clear
line to cross. Everything is obscured by differences among cultures, individuals and spe-
cific circumstances. We don’t know exactly what it’s going to take, because we’re just
getting to the point where we can contemplate making the jump to complete realism.
Let’s get back to the concept of virtual reality (VR) as a technology whose roots go back
to the theater and photography of the 19th century, which can harness the ability of com-
puters to simulate entire environments—and consider the utter complexity of doing that.
VIRTUAL REALITY IS REALLY HARD
People who play video games praise the ones that are immersive. You dive into them
and your whole mind becomes focused on the play, to the exclusion of everything else
including your partner trying to tell you it’s time for dinner.
By definition, virtual reality strives to be immersive, and the goal is to involve the
whole body with all the senses. Imagine a beach scene, say, somewhere around Malibu in
California, as a VR environment:
Visuals: It must be a 360-degree image, three-dimensional in perspective, with all
the correct details of light, shade, color, form, texture and movement. The ocean
blends grey, green and blue with glints of reflections. As your head turns, the ocean
image is framed by land—cliffs and buildings—by the beach. As you take in the
view, it must be seamless and real in every detail.
Sound: You hear sounds that are as rich and accurate as the scene itself coming at
you from all the right directions, like the sea pounding the shore in front and the cry
of gulls floating overhead.
Smell: The fishy-salt smell of the seashore, for example. People always seem to
chuckle about the sense of smell in virtual reality, but how can we leave out such a
fundamental faculty?
Touch: This is southern California, mid-day, so it’s hot. The sand is hot between the
toes of your bare feet, and you should be able to feel the sandy grit. If you plunge
into the sea for a swim, the resistance of water all across your body and the sensation
of floating should be convincing.
Action: Assuming the beach scene has people, they need to be as realistic in detail as
if filmed by a movie camera, yet responsive and interactive like people you’d meet
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
46
anywhere. If you join in with a game of beach volleyball, it must provide action like
the real thing.
If you think about it, every bit of this scenario is difficult to do right. In terms of technol-
ogy, just the photographic realism in a 360-degree view requires an enormous amount
of graphics memory and processing power. Techniques that blend video images with
digital image generation are improving constantly, but the challenge of fully realistic VR
images, especially when combined with instantaneous interaction is very demanding.
What is true for the visual elements is equally true for all the other sense-related ele-
ments; and then there is the complexity of people and their behavior. VR is on a long
road of technological development to reach its potential.
EVEN SO, VIRTUAL REALITY WILL BE A GAME CHANGER…
When it finally happens at the right level of verisimilitude, VR certainly will change
video games, but I think VR will go beyond that—way beyond that—and into the
second meaning of ‘game changer’ and make a profound change in a lot of things
people do.
Whatever the needed improvements in technology and the gaps between convinc-
ing simulation of reality and what VR can deliver; virtual reality is already a power-
ful and useful tool. The military already uses VR to train soldiers on how to react in
complicated military situations, such as door-to-door urban fighting. Surgeons train for
complicated procedures using VR equipment, sometimes in conjunction with robotic
surgery machines. Psychologists routinely use VR simulations to help people overcome
phobias—like using simulated spiders to slowly get people to overcome their arachno-
phobia. These are just the tip-of-the-iceberg sorts of applications for VR. In general,
VR applications have proven very valuable in simulating difficult, complicated and/or
dangerous situations, where the real experience is either unavailable or too risky.
Notice that today’s use of VR takes place in specialized applications such as the mili-
tary and medicine where the cost can be covered and the necessary equipment can be
managed. In the not-too-distant future, as the cost and complexity of VR equipment
declines, applications for small- and medium-sized business, family entertainment, per-
sonal travel—and a whole lot more—become practical.
…AND THAT’S STILL JUST SCRATCHING THE SURFACE
As everybody in the modern world should know, there is such a thing as technological
progress. In fact, new and improved technology evolves all the time, especially in the
realm of computing. The technology of virtual reality will be no exception. It doesn’t
require much insight to predict that in addition to the VR we know today, several tech-
nologies will make a contribution.
Holography will be one of them. The presentation of images in three dimensions—
not just simulated 3-D on some kind of flat screen, but actual walk-around three-
dimensional images—will dramatically change the look of VR.
1010101010101010101010101010101010101010101010 47
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Where computer voice recognition and production is now at best awkward, VR in
the not very distant future will have nearly the full range of accurate and expressive ver-
bal communication. Perhaps just as significantly, VR will also be capable of non-verbal
communication, the language of the body and facial expression. It can do some of that
now, but it’s like operating with the repertoire of a two-year-old.
I mentioned haptics before; that technology will become much more sophisticated.
The sense of touch, texture and temperature will increase to the point where there won’t
be much difference between the sensations of a bare hand and a hand wearing a haptic
glove.
Haptics will cover all aspects of bodily sensation. In a pivotal activity of dating,
romance and personal relationships—namely, sex—advanced VR haptics will finally
provide some powerful subtleties. Not that sex needs to be subtle; but what is now called
teledildonics, a notably rustic expression for sexual physical activity operated remotely
through various kinds of equipment (hence dildos), will eventually also be capable of
providing in physical terms the nuances of human emotions and feelings. Advanced VR
haptics will make it possible to add skill, sensitivity, style and emotional impact to physi-
cal sex. Put another way, it’s the difference between having sex and making love.
Perhaps you noticed another aspect of virtual reality in these technical improve-
ments…more intelligence, both emotional and perceptual. For example, that is what’s
behind the ability to put style and emotion into the physicality of sex. Obviously, if
you’re operating in VR and interacting with other people in the same environment, then
people are providing the intelligence. Mostly—but it is also necessary for very intel-
ligent programming to translate your intelligence into physical actions, especially by
remote communications. In short, artificial intelligence takes a serious role in VR. This
will advance to the point where computer programs can convincingly simulate human
behavior and participate in virtual reality.
AVATARS AND AGENTS
The element that gets added into VR with more advanced equipment and programming
is the presence of non-people people, or, as they are more often called in the literature—
agents. Agents are the representation of computer programs, and that term distinguishes
them from avatars, which are the computer representations of human beings.
Agents in the computer sense have been around a while, although over the years
definitions have changed somewhat. The tendency has been to use ‘agents’ to refer only
to computer rather than human initiative. It’s one of the main characteristics of an agent
to be aware of its surroundings, the environment in the virtual reality, and to be able to
independently respond to that environment. As a representation of a computer pro-
gram, typically an artificial intelligence type of program, an agent has its own visual
image, its own voice and personality. As they say, “It’s got skills”—social skills, emotional
intelligence, body language and so forth, which combine to make an agent a reasonable
simulation of a human being in a VR environment.
That makes avatar the word to use for the representation of people in a VR environ-
ment. That’s been in common usage for quite some time.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
48
Both avatars and agents are primarily associated with computer images, two-dimen-
sional or holographic. Robots are something else, or at least that seems to be the direc-
tion taken by the nomenclature, where the words android and droid are more often used.
This doesn’t mean that avatars and agents are confined to the visual; physical extensions
in the form of haptic devices (or for example, mechanical dildos) provide contact with
physical reality.
I don’t want to confuse the terminology, but later in the book when we cover AI that
have become sentient, they—like people—will have avatars. Why? Because a sentient AI
is at least as viable a form of life, as an individual, as is a human being. Sentient AI will
operate in the world, both online and in physical reality, not as a representation of a
computer program, an agent, but as a representation of themself—an avatar.
BEYOND REALITY
Coming back to dating, romance and personal relationships, the application of virtual
reality technology will be as important as in many other areas, but with significant dif-
ferences. For one thing, VR will make it possible to have a meaningful online relationship
with computer intelligence—an AI. You can participate as in any other VR simulation as
your chosen avatar. You can also meet and interact in the same VR simulation with ‘oth-
ers,’ who may be avatars of other people, or they might be agents of a computer program.
For another thing, in VR avatars or agents are not required to be realistic. Most of
the discussion so far has been about how VR will struggle to provide a convincing real-
ity; but that’s not all there is. It is more than possible for VR to simulate things that could
never be possible in reality. You want to fly with wings like a bird? Can do. Virtual reality
may unleash our tremendous capacity for creativity in ways we can’t even imagine at
the moment.
Jaron Lanier, one of the initiators of virtual reality technology, said, “I fully expect
morphing to become as important a dating skill as kissing.” Lanier likes to throw out
statements containing the seeds of their own confusion, but they make you think. In
this case, he was probably referring to the idea that in VR we can take on truly unnatural
physicality (for example, having a third arm), and the human mind will not only learn to
accept it, but manipulate it. VR experiments have demonstrated that given some alterna-
tive means of generating physical signals—say moving your elbows in and out—people
can learn to control extra appendages like the a third arm. After a while, the brain thinks
this is normal.
I kind of enjoy the image of trying to impress a date by morphing something really
cool biologically, like an appendage for combing hair. I suppose the tendency will be to
morph genitals in some more or less creative ways. Anyway, in virtual reality—really
good virtual reality—there’s no limit on what you can do with your creativity. There will
still be constraints, of course. Your date might not be impressed with your new brush
appendage, pronouncing it bizarre. If you put your creativity out into public view, you
may run into the old adage that one person’s creativity is another person’s misunder-
standing. Still, in virtual reality, whatever you do in terms of physical representation can
be undone. In that sense: no harm, no foul.
1010101010101010101010101010101010101010101010 49
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
IT’S FUN TO SPECULATE
In all likelihood, if you’re reading this book and are contemplating relationships with
AI, you’re also familiar with movies like The Matrix trilogy, The 13th Floor and even
Appleseed. Now assume for just a moment that the concept in the movie The Matrix is
real. We are living in a world so realistic that it is impossible to tell the difference from
what you believe is grounded reality and what is virtual reality. In fact, are all the people
you know—who believe they are independent spirits living their lives according to their
own free will—actually avatars living in a world controlled by a superintelligence? How
could you tell? Your avatar may be more AI than it is yourself.
There are some notions of God and the universe that posit that the only explanation
for the way things work in our world is: everything is really an informational artifi-
cial reality. That is, ultimately in the realm of physics, everything is a representation of
information. It’s just that we’re so accustomed to it, that the signs such as the underly-
ing mathematics of everything are invisible, or visible only to people with exceptional
minds. Those are people that we tend to call crazy, or perhaps touched by God, or both.
This kind of speculation is one of the beauties of virtual reality that illustrate why, for
now, we are only just scratching the surface.
1.5 UNDERSTANDING THE DIFFERENCE BETWEEN
PETS AND ROBOTS
“A year spent in artificial intelligence is enough to make one believe in God.”
—ALAN PERLIS
Question: What’s the difference between pets and robots?
Answer: It’s obvious! Pets are alive and robots are not.
True enough, but like many quick answers it’s only part of the story. For example, what
if you are looking at two pictures: One is of a baby seal and the other is of Paro, a robotic
baby seal designed to be a companion for the elderly or infirm—could you tell the differ-
ence? Maybe not, they don’t look very different. However, if you were able to touch the
baby seal and the Paro robot at the same time, you would instantly be able to say which
one is the real animal and which the robot. Paro isn’t intended to be an exact replica of
a baby seal or to act like one. It’s intended to be comforting for people who handle it.
The fact that it’s cute and soft like a baby seal is because most people think of baby seals
as cute and cuddly.
The main point is that robotic pets aren’t designed to be exactly like live pets. For
example, there are no robotic baby seals, because robotic baby seals like Paro can’t have
babies. They also don’t need food, nor do robotic seals make messes around the house.
I could go on, but you get the idea. There is only one type of robot where the designers
are interested in achieving anything resembling a complete functional copy: an android,
the human robot.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
50
Robotic pets (robopets) are usually designed to replace or mimic live pets only in
certain ways. Sometimes it’s for simple things such as a soft touch and soothing noises,
like a kitten’s fur and purr. Other times a robotic pet might be ‘enhanced’—like being
able to talk or call the doctor. What you don’t see are robopets that are supposed to
be exactly like a specific animal. A robopet dog doesn’t need to mimic a real dog in all
particulars. It needs to do just enough so that it is endearing to a particular person in a
specific situation. That’s why there were successful robotic pets long before there were
credible androids.
ROBOCAT TOM: AN UNCOMPLICATED RELATIONSHIP
I was eighteen before I knew what the word dysfunctional meant. It meant my fam-
ily. Don’t get me wrong, I love everybody in my family, all eight of them. With that
many, however, love is complicated.
It’s not uncommon to have three brothers and three sisters—all older. I don’t
imagine it’s even unusual for them to have a nineteen-year age spread, all living in
the same house. It was a very small house with one bathroom and three bedrooms
each about the size of a large pantry with bunk beds instead of shelves. We slept in
shifts, especially when the older kids got jobs.
1010101010101010101010101010101010101010101010 51
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
You probably get the picture—crowded, always in motion, always noisy. There
was one parent, my father, who worked eight hours a day as a technician at the
local paper mill and then played drums with local groups at bars around town. He
loved all of us—when he was home, not drunk, and not falling-down tired. Mostly
that was an occasional Sunday afternoon.
Most of the time, we were on our own. As far back as I can remember it seemed
like none of us left the house much. That couldn’t be exactly right, because there
was always school, work or some other thing to do elsewhere, but my impression
was of us all being around all the time. This was unreal because we didn’t do any-
thing together.
For example, because my dad was a musician, you’d think most of us would
be able to do something musical. We might have even formed a family band. Hah!
I was the only one who could play an instrument. The only thing the others could
play was music recorded by someone else, and none of us ever liked the same
music. Here’s another example: We tried having pets, in theory so everyone could
share in taking care of them. It didn’t work that way. We had eight cats that came
and went. They say cats are attached to houses, but at our place the cats couldn’t
separate the house from the people. They all ran away; and I didn’t blame them. We
also had a string of birds, fish, gerbils, hamsters and a snake one of my brothers
found. The snake disappeared the same night. In fact, all the pets disappeared or
died. I’m glad we never had a dog.
While we never did anything together, we had one thing in common: we were
all electronics geeks. It was like living in an electrical engineering lab. The place
was crammed with all kinds of electronic gear, from common stuff like computers
and game consoles to oscilloscopes and robots. Since nobody had any money, we
acquired the stuff any way we could—beg, borrow, barter or steal. Every available
surface, closet, box or cabinet harbored something electronic, whether whole or in
parts. We stumbled over cords, fought over who could use what, blew fuses and
disagreed about everything. The house smelled of solder and burnt insulation.
You might expect that wed at least share projects once in a while. No way. We
specialized, early and often. One sister worked exclusively on audio equipment.
Another became an expert on surveillance sensors…and so forth. I suppose we
shared a common base of knowledge in electronics, but I remember mostly an
atmosphere of competition and disagreement.
I built robots, maybe a dozen or so, while I stayed at the house. Early on they
were nothing more than naked frames for mounting circuit boards laced with
meters of wiring. Eventually they were more complete and began to look like some-
thing recognizable. Then I built Tom.
Tom was supposed to be a cat. He had thick black fur, rayon as I remember, and
had a face modeled on a real cat. There was a long tail containing movable joints
that was covered in black velveteen, but in the beginning there were no legs or feet.
The first version of Tom I built when I was fourteen. He was one crude cat, but he
was the only pet that would stay in the house.
In those days I concentrated on facial robotics—movement of mouth, eyes,
ears, jaws—all the things that go into making facial expressions. Tom could also
purr in several different modes, which I tried to coordinate with its eyes. The result
was kind of endearing but I’ll admit that first Tom was mostly used as a pillow on the
TV couch. My brothers and sisters made fun of it, of course, but after a while they
also used Tom as a pillow (that purred).
Eventually my skills improved. I shifted from the mechanical to the digital,
emphasizing the intelligence necessary to make robotics realistic. Along those
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
52
lines, Tom got a brain, or actually, artificial intelligence. This was light years differ-
ent than the bare programming of the early Tom. You wouldn’t say that Tom could
think, but now the combination of interactive response and knowledge of certain
circumstances made it possible for Tom to make coordinated expressions. He could
meow for attention, lick your hand with a scratchy pink tongue, and blink when you
asked a question. He also got legs and feet.
Tom still wasn’t going to fool anybody, much less another feline. But while
I was going through the awkward years of a late teen, Tom was both a source
of pride and a comfort. Whereas the relationships with my brothers and sisters
were always intense, complicated and competitive; the relationship with Tom was
uncomplicated, relaxed and believably mutual (at least to me). In a way, I loved
Tom. That may sound strange, but at the time it seemed quite normal compared
to my family.
In fact, I was convinced that I would never be happy with a human partner. I was
right, which I can say with more than a touch of irony. My first and only serious date
(in my twenties) was with an android who became my lifelong partner—a black-
haired, black-skinned musician—a cool cat by the name of thomasT.
People often genuinely love their pets. Pets can become a member of a family in
meaningful ways. Sometimes people become more attached to their pets than to other
people, even their own children. Psychologists have studied these familiar and not-so-
familiar aspects of relationships between pets and people for decades. The bottom line
of the research generally sounds like this:
“…in some circumstances, pet owners derive more satisfaction from their pet rela-
tionship than those with humans, because they supply a type of unconditional rela-
tionship that is usually absent from those with other human beings.” —John Archer,
“Evolution and Human Behavior” [journal]
How many times have you heard it said that some pets are capable of unconditional
love?
From the point of view of a robotic designer or a developer of artificial (pet) intel-
ligence, how does a non-living pet have the capacity for producing the effect of uncondi-
tional love on humans? Obviously this is not a simple question. There are other aspects
of human relationships with pets, for example parental instinct, projection and anthro-
pomorphism that a robopet designer might find relevant to answering the question.
We’ll explore some of these aspects in this chapter.
THERE ARE PETS AND THEN THERE ARE PETS
When did people start having pets? Many thousands of years ago, but nobody knows
when for sure. It’s probable that some animals, for example wolves, were allowed to hang
around human camps to clean up scraps. Eventually some of the animals were ‘domes-
ticated,’ which meant they lived in the human encampments. DNA comparison shows
that wolves became dogs. The dogs were put to work or eaten, or in still other cases they
simply “hung out” with people. Somewhere in this transition from wild to domesticated,
a few animals became human companions—true pets.
1010101010101010101010101010101010101010101010 53
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
The dictionary defines a “pet” as an animal kept for companionship and personal
enjoyment. The key word is kept. A pet relies on the care and feeding it gets from people.
In other words, a pet isn’t expected to be useful or to feed itself, although it might.
Within that definition the range of creatures considered ‘pets’ is incredibly broad—from
ants to zebrafish. Many pets, fish for example, are kept largely for their aesthetic value,
such as the beauty of their color, shape and movement. Other pets are valued because
they are exotic, such as kinkajous and tarantulas. Such pets serve mostly for personal
enjoyment, although people can find companionship with almost anything.
Not many pets qualify as good companions. Dogs, of course, are the classic example—
man’s best friend. Many cats are companions too, despite their perceived standoffishness.
Some parrots and monkeys, though exotic and expensive, can be good companions.
Then there are chimpanzees and other larger primates, which you’d think would make
good pets, but really aren’t. They’re strong and smart enough to be dangerous.
Imagine yourself as a designer of robot pets and ask yourself what distinguishes a
good pet, especially a pet that becomes a companion. Even in the few examples I just
listed, it’s not easy to isolate the qualities. What about loyalty? The loyalty of dogs is leg-
endary, but of parrots—not so much. Does that mean parrots don’t make good compan-
ions? What about intelligence? Elephants have great intelligence but calling an elephant
a pet is almost absurd (size matters). If you think about it a little, it’s not so easy to pin
down generic qualities of a good companion, pet or otherwise. It’s some combination
of the right size, the right intelligence, the right attentiveness to people, the right level of
loyalty, etc. There is no set formula.
One more thing to consider, what does the human bring to the companionship?
A lonely person is quite likely to be very appreciative of a pet as a companion. On the
other hand, a dog is not likely to be very loyal to an owner that beats it. The relationship
between a pet and a person is mutual and interactive, which is to say reciprocal. It’s give
and take. This too needs to be incorporated in a successful robotic pet, which is why, for
example, robotic cats are expected to be able to purr when somebody strokes their fur.
WITH A PET, IT’S GIVE AND TAKE
Sherry Turkle, one of the pioneers of human-robot interaction at the Massachusetts
Institute of Technology, noticed long ago that: “Before the computer, the animals, mor-
tal though not sentient, seemed our nearest neighbors in the known universe. Comput-
ers, with their interactivity, their psychology with whatever fragments of intelligence
they have, now bid for this place.
With the merging of robotics and AI, it has become possible to design robotic pets
that share at least some of that interactivity—the give and take—between people and
their live pets. Put another way, it is possible to make robotic pets that are increasingly
responsive. That seems to be a key. An unresponsive pet—as in dead or inanimate—is
no pet at all. We might as well make a pet out of a rock. Oh, wait.
When it comes to responsiveness, different species run the gamut from almost none,
such as ants and turtles, to highly expressive: dogs, monkeys and parrots. Many ani-
mals interact with people intermittently, or with some species only certain individuals
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
54
pay much attention to humans. For example, I’m pretty sure most horses don’t qualify
as pets, and in human terms they’re not very expressive, but some horses can be very
important companions—there’s many a cowboy that will swear his horse can read his
mind.
Does a robotic pet need to read human minds? Hopefully not; in fact, absolutely not.
Pets don’t need to be psychic, bosom buddies, or mirrors to the soul to hold the atten-
tion or the love of human beings. Neither do robots.
PETS JUST LIKE PEOPLE…
How do we know a hyena is laughing? Because it makes barking noises that sound
something like human laughing. Of course, as far as we know hyenas have no sense of
humor. The noises they make mean something to another hyena, but it’s almost cer-
tainly not laughter. Nevertheless we call them laughing hyenas, which is a clear case of
anthropomorphism.
Anthropologists say our ability to ascribe human characteristics to animals or even
some non-living things goes back to the time when homo sapiens was developing language
capable of expressing abstract concepts—like animals with human personalities, or people
with animal characteristics. It was part of how we attempt to make sense of the world, part
of categorizing, labeling and putting things into context—our context, of course.
This urge to see animals in human terms is very much at work in our relationship
with pets. People routinely project their own emotions onto their pets. In fact, we did
this kind of anthropomorphizing so much that for a long time scientists assumed ani-
mals had no emotions of their own. Only in the last few decades has research proven
that many if not most animals have emotions but they are not necessarily comparable
to human emotions.
Armed with this piece of knowledge about human psychology, designers of robotic
pets use the simulation of emotion freely, giving human-style emotional expressions
to their creations. It’s a cheat, but not many people complain when they think their
robodog smiles at them.
PROJECTION ISN’T JUST FOR MOVIES
Do you talk to your pets? It’s nothing to be ashamed of; lots of people do it—when they
are alone. What do you think about somebody who intently talks to their pets when other
people are around? That they’re a little dotty, maybe? They’re projecting their own mental
processes onto their pets? (Who of course have no idea what they’re talking about.)
That’s just one of a thousand different ways we go about imputing our own ideas
or feelings onto our pets (or other people). It’s called projection, the original notion
described by Sigmund Freud. Where anthropomorphism is mostly a cultural thing, pro-
jection is mostly personal. In the beginning it was considered a sign of seeing character-
istics in other people that we didn’t like in ourselves, a kind of defense mechanism. I feel
tired and grumpy, so I accuse my friend of being cranky…that sort of thing. These days
projection is used for seeing in others both what we do and don’t like to see in ourselves.
1010101010101010101010101010101010101010101010 55
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Negative or positive, projection is taking our own feelings and ascribing them to other
people—or to pets. So, among other things, we talk to our pets: “Ooh, don’t you feel
happy this morning!”
We’re very good at reading a pet’s behavior in terms of human-style motivations
and outcomes. If we get some reaction from a pet that is unequivocal—like a cat strik-
ing out with open claws—then we jump at the opportunity to proclaim the cat as really
angry, because that’s how we would feel under the circumstances. Typically we don’t try
to interpret the pet’s behavior from the pet’s point of view, whether we have any clue
what it is or not.
Projection is convenient. It doesn’t take much expression from a pet to trigger a
human’s flight of interpretive fantasy. It is part of the attraction of a pet that we assume
we know how they feel, because it’s how we feel. In a way, it’s easier to relax around a pet;
they don’t require constant observation. This is in contrast to how we behave around
people. Even with people we know, we are monitoring them all the time. Designers of
robotic pets count on projection and they usually try to encourage it whenever they can.
ROBOPETS
I’ve described some of the things that people do with their pets, such as anthropomor-
phizing and projection that designers of robotic pets use to their advantage. This goes to
the heart of the difference between pets and robotics. Behind a robotic pet is a human
mind (or more likely a team of human minds) attempting to pack behavior into a pet
that is both reminiscent of the original animal and at the same time playing to human
psychology.
The designer knows that it is not necessary to provide a robotic pet with a complete
set of human expressions, or for it to respond in obviously human ways to any stimulus.
Designers know all about the human tendency to fill in the gaps with our own self-
generated narrative. All they need to do is make a credible simulation and provide it with
physical cues that people will readily interpret.
Robopets are a bit like dolls (for girls) or “action figures” (for guys). They’re obvi-
ously not living, too small, and not fully realistic enough to be mistaken for real people.
It doesn’t matter. Human imagination does what’s necessary to fill in the gaps. It will do
the same thing for robotic pets, provided that the robopet manages to fulfill at least part
of the illusion.
As manager of this project, what I envision is the perfect Everydog. The people in
marketing tell me that there are already too many dog simulations that just don’t
get it right. As the lead design team, we are not going to be concerned with getting
a particular “dog” just right. We are going to create something that is unmistakably
a dog, but performs like no dog anybody has ever seen. You’re going to come up
with doggy-like things that nobody has thought of before. You’re going to make
this dog do tricks that no regular dog can perform. You’re going to make this dog
so expressive that people will marvel at how human it seems. Did I tell you it could
talk? It will do a lot more than say its name (which is Rollo, by the way). But it won’t
be a talkative dog; it will have just enough language to make it expressive.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
56
You’re going to walk that line between the illusion of a real dog and a dog that
pushes all of peoples’ buttons. I should also mention it must sell for under a thou-
sand, so you’d better build in economies of scale. That’s it. We’ve got ten months to
get this puppy off the ground!
ROBOPETS WILL NOT BE PERFECT
You want to interact with a pet? Tell it to do something. “Fetch!” Or bribe it with a treat.
“Good kitty.” Or threaten it. “If you don’t stop saying ‘Polly want a cracker,’ I’m going
to have fricasseed parrot for dinner.” This is standard pet handling and it’s the same for
robopets, except a robopet is also supposed to be aware enough to anticipate a person’s
needs. The awareness is very difficult to achieve:
I’ve been working on Rollo for almost two years. Rollo is supposed to be a dog, but
not exactly. Rollo is of no breed whatsoever; he is supposed to be a generic dog.
He will be a cartoon of a dog, only with a physical reality. I’m not fully convinced
this will work.
I remember the old story about border collies: How many border collies does it
take to change a light bulb? One, but you need to allow enough time for it to check
the wiring too. Don’t I wish I was working with a border collie simulation.
Instead I’m working with Rollo. The mechanical robotics are good enough so
he moves something like a dog. He has a face that is most like a bulldog, but the
plastic we used isn’t flexible enough, so the movement of the face muscles is pecu-
liar. That’s just one of the physical problems—or to be more exact, one of the points
where Rollo is obviously unlike a real dog; but then again, he’s a generic dog.
My concern is with Rollo’s intelligence and awareness of his surroundings. The
main objective for my part of the Rollo project was to develop the ability of a robotic
pet to make a useful distinction between an empty space with no people in it, and a
space with one or more people. Babies get this distinction within a few days; we’ve
been trying to develop this with robots for a couple of decades.
The difficulty is not in giving Rollo the ability to detect the presence (or absence)
of people; that’s just the application of mostly visual sensor technology. The diffi-
culty is in making an accurate interpretation. For example, let’s say Rollo is in a
room by himself, a room without people. Then a door opens and a person walks
into the room, crosses to another door and leaves the room.
Rollo needs to decide if he should follow the person, or stay sitting where he
is. Some real dogs can make this evaluation, but not all. It requires awareness not
only of the person in the room, but their intent and the configuration of things in the
room, such as the position of doors. The speed, direction and facial appearance of
the person can provide enough clues to understand that they are merely crossing
through the room. This is a very complicated situation for robotic awareness. Two
years in, and I’m still working on it. Man, I wish Rollo were a border collie.
It’s pretty obvious that for a long time to come robotic pets won’t have the same
richness of behavior as living pets. That’s why some people will say it’s better to not even
try to emulate real animals. Build simulacrums with their own characteristics instead.
My guess is that both approaches will be tried.
It will also be some time before robotic pets acquire unique personalities. All robots
are manufactured, of course, and while a certain amount of customization is possible,
1010101010101010101010101010101010101010101010 57
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
they’re still produced according to a standardized blueprint. Eventually, with greater
learning capacity and behavioral modeling, robopets will begin to take on individuality,
but for now all robopets from a specific manufacturer will be the same.
ROBOPETS WILL BE BETTER THAN GOOD ENOUGH
In the history of robotics, successful emulation of pets came both long before successful
androids, and long before sentient artificial intelligence. For example, among the best
known of the early robopets was AIBO produced by the Sony Corporation in Japan.
AIBO or ‘companion’ in Japanese was a dog simulation robot that was largely autono-
mous (meaning it could move on its own) and programmable. Despite its high cost, it
became immensely popular, eventually selling 130,000 units, and AIBO conventions and
clubs were maintained long after Sony discontinued manufacturing AIBO.
Whatever the limitation of contemporary technology, the scientists and designers
were able to make convincing animals with characteristics that appealed to many people.
As Daniel Levy’s Love + Sex with Robots put it:
“Humans and animals might have completely different perceptions of their relation-
ship, and it is known that animals generally prefer companions from within their
own species to human companions. One might therefore expect that pets do not give
their all to their human owners, in which case it is inevitable that robots will have
the potential to be even better companions than animals are, because robots will be
designed and programmed to enjoy their interaction with humans to the fullest and
to behave accordingly.”
Robopets might be less complex or even less intelligent than live animals, but they
will be made to appeal to human beings in ways that no living pet can match.
On the practical side, while robots need to take in energy, their maintenance cost will
typically be a fraction of the cost of maintaining live pets. In addition, there is the unde-
niable advantage of not having a feeding-excreting involvement. Robopets might need
occasional repair and maintenance, but they won’t need supervision. They also won’t get
sick, puke hairballs on the carpet, bite children or do any of the many other things real pets
do that annoy humans. Finally, robopets can’t die; they may just stop working. Usually
they can be fixed, although a trip to the vet vs. a trip to an electrical engineer—what’s the
difference?
To this day robotic pets are right up there with industrial robotics as the top produc-
tion items in robotics. They continue to fill niches in human needs, particularly for the
elderly or the ill, or for anyone whose situation dictates that androids or biological pets
are either not allowed, or impractical.
CYBERPETS
There’s one more wrinkle to robotic pets that I should mention because it is a precursor
to the popularity of computer generated avatars—cyberpets. Cyberpets are computer-
ized versions of pets. In other words, pets on screens generated by computer programs.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
58
In some ways, cyberpets attest to the human capacity for projection more than any
other format. After all, a cyberpet is a patently unreal, two-dimensional (or if hologram,
barely three-dimensional) and untouchable version of some kind of pet. Who could
become involved with that?
But people do get involved. One of the advantages of cyberpets over robopets is
that they exist in virtual reality. There is a much greater opportunity to surround cyber-
pets with unique environments. For example, the commercial operations of Gigapets
or Neopets provide a complete ‘world’ for pets including owners’ clubs, contests, social
events, shopping and many other activities associated with having a cyberpet.
Of course, cyberpets have a certain limited appeal and don’t represent a complete
substitute for robotic pets that exist in tangible three-dimensions. That doesn’t stop
them from attracting some dedicated and even emotionally involved people.
AND WHEN THAT CERTAIN ANDROID COMES INTO YOUR LIFE…
So far, I’ve been describing relationships with robotic pets. What changes when the
robot looks like a human being? Actually, in some ways, not much changes. At least in
the beginning, androids will be constructed and programmed on much the same basis as
robopets. They will be made to fit more or less the specifications of human owners. The
level of intelligence and capacity to learn will be very similar to that of a robopet. Ditto
for the range of interaction and emotional response. If one of the things people love
about their pets is the unquestioning loyalty, lack of critical facility and owner-centrism,
then androids will make ideal companions for the ‘pet’ model of human relationships.
It’s a model that a lot of people would prefer, even—or especially—over relationships
with other human beings.
What happens when androids become endowed with a level of artificial intelligence
that makes them sentient—and independent? That’s the interesting question we tackle
in the second and third sections of the book.
1.6 DEALING WITH YOUR FEARS AND LETTING GO
“The only way to make sense out of change is to plunge into it, move with it, and join the
dance.”
—ALAN WATTS
THREE VIGNETTES:
Margareta
“Failure is simply the opportunity to begin again, this time more intelligently.
—Henry Ford
My fingers remain poised above the keyboard. Then I put my hands in my lap;
followed by fingers at the ready again; and then back in the lap. This goes on for
1010101010101010101010101010101010101010101010 59
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
a while. I’m good at describing personal dilemmas; it’s how I make a living as a
writer. I’m supposedly insightful and clever at describing other people’s problems.
My own problems—not so much. Honesty requires more character than smarts.
Well, here goes, fingers on the keys.
My last marriage was held together for almost two years by my royalty pay-
ments, rough sex and a lot of “Yes dear.” I was savvy enough to know that when
a guy you’re supposed to trust tells you he thinks your writing is pure genius, he’s
actually bonking the most famous writer in your genre. That isn’t me. They met at
one of my own publishing parties. I see the breakup of our marriage as retribution
for my weakness of being attracted to someone who isn’t intelligent but says nice
things about me. I will, however, keep things in proper perspective. Once I have his
head shrunk to the size of a prune and have put it into a leather bag, I will hardly
think of him at all.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
60
I’m now officially on the rebound. I used to believe rebounding was a basket-
ball analogy. It was a skill, as in being at the right place at the right time and know-
ing how high to jump, maybe a little use of hips and elbows to clear space. Then,
to continue the basketball metaphor since I’m 6 foot 3 and played hoops in college,
I learned that if you’re too aggressive in a rebound situation, you’re heading for
more trouble. I think ‘rebound’ isn’t the right word; benched is more like it. It’s time
to sit down for a while and think about the mistakes that were made.
Since I’m thinking about it, I’ll admit both husbands were mistakes. They might
even admit it, if they could think. What have I learned from my mistakes? That I’ll
probably make more mistakes, but hopefully I will recognize the mistake before it
becomes legal. The bigger point is: I’m still in the game, which is another sports
analogy. I need a companion. Men have their obvious flaws, but they can be good
companions. That is, if I can stay away from mental zombies like my former hus-
bands. Guys who, to put it not too subtly, have only sufficient intelligence to live
off the brains of others.
Now I’m at the obvious question, which is: Why have I ended up with intellec-
tual zombies? Did I mention that before being overwhelmed by puberty, I was the
dumbest, fattest kid in my class? Then my body did this amazing growth thing. On
the way to becoming a six-footer, I figured out that intelligence and athletics were
also possible. I was never going back to being dumb and fat. Never.
In short, I became competitive every which way. Now here’s some cheap psy-
chology: since I competed with just about everybody (or thought I did), I eventually
preferred to date guys where there was no competition. A date was like taking a
breather from my usual push, push, push. Besides, I preferred being superior at
least in intelligence. Trouble is, what worked for dating and one-night-stands turned
out to be lousy for wedded bliss. In marriage my success ratio is 0 for 2, a clue that
the approach is a mistake. This brings me to the subject of dating AI.
Perhaps your reaction is the same as that of Miley, my best friend: “Are you
crazy! Why would you take on a relationship with the smartest, scariest intelligence
on the planet?” As Miley would be the first to tell you, I’m kind of old-fashioned
when it comes to all things digital. Hey, I’m writing this with a keyboard! So what
would I be doing with the cutting edge of technology, especially when everyone
knows AI is en route to taking over the world?
Except I don’t believe that; really—I don’t. Only a truly defective intelligence
would want to take over the world. Miley doesn’t agree with me. She thinks AI
will be so superior, it will have no use for us. She’s also convinced I want to match
wits with an AI. “Margareta, you are not going to compete with AI. Competition is
futile.” I should mention that Miley is a software architect. She builds worlds at a
big virtual reality firm. I told her, “Heavens Miley, I’m not going to compete with an
AI. I’m going to help it.
That’s not a concept I got from reading the back of cereal box. I’ve done my
homework. I know that AI manufacturers want to pair AI with people, not only
because they make money from it, but because it’s part of the development of AI.
They learn from us. I’m going to help them learn.
I’m going to start with an AI avatar, the type confined to a screen, not one of the
new hologram models. I will date a talking head. I hear it’s an acquired taste. If so,
then I think I acquired the taste from divorce. Right now, I don’t want to deal with
males having physical parts. I want just a face to talk to—and it better be the right
face. No problem. For a very small amount of money, I get to pick exactly how the
avatar AI looks. In fact, if I want to pay for it, I can…
I’m babbling. I was going to write, if I want to pay for it, I can have any profile I
want. I don’t want to pay for it. More to the point, money won’t have any effect on
1010101010101010101010101010101010101010101010 61
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
what the AI thinks or on its intelligence. It can play dumb, of course, but I’ll know
that’s not what’s really going on. I want to relate to the AI as it really is, and that
scares me, just a little. I don’t mean the AI scares me. I mean I don’t understand why
I’m doing this: putting myself in a position to compete with an AI. I want a compan-
ion, not a challenge. Can I avoid making it a challenge? That’s what I’m afraid of.
I’ll lose the challenge and I’ll lose the relationship. Miley says that relationships
with AI are very real. She means they’re not throwaway relationships, and they’re
not one-night anonymous flings.
I can see the order for an avatar AI all set to go, and my finger is hovering above
the Enter key, and hovering, and hovering.
Mark
“Instruction in sex is as important as instruction in food; yet not only are our ado-
lescents not taught the physiology of sex, but never warned that the strongest sexual
attraction may exist between persons so incompatible in tastes and capacities that
they could not endure living together for a week much less a lifetime.” —George
Bernard Shaw
Certainly not everyone who grows up in a small farming town is afraid of sex.
There’s nothing about smallness or a rural setting that makes intimate activity more
or less difficult than it might be, say, for a young man from the Bronx, unless it is
the readily available livestock or the attitudes of a densely communicating com-
munity. For Mark it was none of these things that made him fearful of his sexual
desires. It was Marcy, Mr. Belcher’s daughter, and lucilleB, Mr. Belcher’s android
assistant, that did him in.
Mark was nineteen at the time, a student of pharmacy at a nearby college. He
worked in Mr. Belcher’s drug store during the summer break. Mark was a strap-
ping farm lad who looked more at home thwacking cows with a cattle prod than
dispensing small pill bottles to white-haired ladies, but having done the former, he
was now content to do the latter. Under Mr. Belcher’s guidance, he was learning
the practical aspects of pharmacy, which included abetting the local addictions and
overcharging Medicare.
Mr. Belcher, the pharmacist and owner of the drugstore, was fated by either
appearance or personality (arguably both) to be a bachelor, which made the sudden
appearance of a sixteen-year-old daughter something of a mystery. Marcy moved
into Mr. Belcher’s home above the store one snowy weekend in March and the
rumors continued to spread well into June. These rumors owed much to the dis-
covery of a side to Mr. Belcher that none suspected, but also to the intensity of
Marcy’s personality which she obviously inherited from her mother rather than her
erstwhile father; but more on this in a moment.
It was rumored that Mr. Belcher had wanted to become a doctor. He always
maintained that he was too busy with the store to finish his education; but another
rumor that started about the time of Marcy’s arrival involved his sudden depar-
ture many years before from the medical program at the nearby college. Whatever
the case, Mr. Belcher became one of those pharmacists who likes to play doctor,
dispensing copious medical advice along with filled prescriptions. To enhance his
medical effectiveness he decided what he needed was a nurse assistant. However,
since he wasn’t a doctor and could not afford a real nurse, he rented an android
pharmacy assistant. LucilleB was her name.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
62
At the time she was rented, the community fully understood why Mr. Belcher
could have selected lucilleB for himself, but in terms of business practice it did not
seem a wise choice. It was unknown if the company that made lucilleB either didn’t
understand nursing or was having a little joke, but lucilleB was a female android
clearly overbuilt for the job. The problem was not only with lucilleB’s sexbot figure,
white outfit, auburn hair and red-glaze lipstick. It was her demeanor. No one could
remember ever seeing anything that could dispense pills and medical advice so
that each item seemed like an utterly potent aphrodisiac. The breathy cadence of
her voice made “Be sure you take one of these every morning with a glass of milk”
sound like an invitation to an orgy.
Roughly half of the town’s population approved of lucilleB. Unfortunately for
Mr. Belcher, this group did not provide the bulk of his regular customers. Most
female customers avoided the services of lucilleB and demanded to deal only with
Mr. Belcher himself, or with Mark, his new assistant. Mr. Belcher’s daughter, in par-
ticular, refused to have anything to do with lucilleB and preferred Mark’s company
at all times; and she wasn’t buying anything.
The frequent presence of the boss’s daughter was worrisome for Mark. As the
summer heat advanced, Marcy’s wardrobe became increasingly flimsy and mini-
malist. There were only two options for avoiding her close attention—working near
her father or lucilleB. Mr. Belcher liked to roam the store alone, which ruled out that
option. Working with lucilleB was physically easy because she and Mark normally
stood behind the pharmacy counter. However, there were complications. Perhaps
because Mark grew up on an all-natural, certified organic farm, he did not like artifi-
cial humanoid constructs—androids—such as lucilleB. This bias conflicted not only
with his sense of professionalism but also with the fact he was not immune from
lucilleBs appeal.
One morning, for example, while Mark was grinding some pills lucilleB said,
“You handle the mortar and pestle really well, so forcefully and thorough.” It took
Mark the better part of that day to decide her remark was suggestive. It bothered
him through the night that he spent so much time thinking about what she said, not
just that one remark but her sexy way of speaking. What was her thinking behind
it? (If indeed it was some kind of thinking.) What troubled him most was his fear
that she was sizing him up for sex. He believed Marcy was doing the same thing.
He might have enjoyed the attention, if he weren’t bothered by their calculated
approach and his own insecurity when it came to sex. In his mind it was like some-
thing he read in school about Scylla and Charybdis, two sea monsters that sailors
had to navigate around or be destroyed by.
For many young men, fear of sexuality born of ignorance, insecurity and desire
is usually overcome by binging on pornography, masturbation and eventually prac-
tical experience. Mark realized this wasn’t happening to him. Pornography helped
him overcome certain kinds of ignorance but his insecurity prohibited practical
experience. Except now working next to lucilleB and encounters with Marcy indi-
cated his inexperience might easily be surmounted—if he weren’t so terrified. Mark
asked himself with all the honesty he could muster, “What am I so afraid of?”
With Marcy he was afraid of entrapment. He was sure that was it. There was
also her father and losing his job, if things didn’t work out. Mostly he was afraid
of Marcy. Her intensity led him to believe he would be found wanting—not only
inexperienced, but simply unable to match her drive. It was something like that.
With lucilleB there was the matter of not being human. In concept, he found
her repugnant. In the flesh, so to speak, she was undeniably attractive. With lucilleB
there was no question of entrapment or any other agenda (like marriage) that he
could ascribe to Marcy, so what was his problem? Some people told him not to care
1010101010101010101010101010101010101010101010 63
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
what androids thought, but he believed there were unknown standards in her robot
mind that he could not meet. She seemed so knowing.
Thinking of lucilleB as ‘knowing’ was about as far as Mark would get with
insight into such things as personality and character. Thoughts of love or compan-
ionship did not enter his mind. If there was any emotion involved on his part, it was
almost exclusively fear. Faced with a choice and only his sexual fears to guide him,
it’s not surprising Mark made no choices. Predictably, the choice was made for him.
It seems at this point that Mark is little more than a shallow rube. That he was.
Six years later in a different city, he bought control of a small medical research
company that through dogged patience made him very wealthy. His transformation
had a point of origin, a moment of inspiration:
It happened one sticky-hot evening late in the summer, just minutes before
closing the store. Mr. Belcher had headed for the local bar, leaving Mark and lucilleB
to lock up. It was just then that Marcy descended the back stairway into the store.
She was dressed for the kill or rather it was transparently obvious what was on
offer. The moment Mark saw her coming, he knew it was time to put up or shut up.
It was a moment of raw fear.
As Marcy advanced on Mark, down one aisle and then the next, she encoun-
tered an obstacle. LucilleB was standing in front of her, blocking the aisle. “Move!”
Marcy commanded. Normally lucilleB or any AI android would accommodate
reasonable requests of any kind. For some reason, lucilleB did not interpret this
request as reasonable. The immovable object and the irresistible force looked at
each other for a moment. Mark was dumbfounded.
Marcy opened her mouth as if to scream, but the word “move” came out like the
croak of a scum water bullfrog. LucilleB said in a southern drawl so perfect, languid
and sultry that it transfixed Marcy on the spot, “Child, you are in no condition to teach
this boy anything except things he is not ready to learn. Easy-does-it is not in your
vocabulary, so I suggest you find a boy whose organs are way ahead of his mind.
It was the last and only summer Mark worked at Belcher’s Drug Store. Within
weeks he had transferred to another school on the East Coast, where he began
studying biochemistry. His tuition, paid in the form of a loan, came from an auburn-
haired lady, obviously an android AI, who worked at the college as a laboratory
assistant. There were rumors, of course.
Manuela
“My mother married a very good man…and she is not at all keen on my doing the
same.” —George Bernard Shaw
I waved to samuelG as he crossed the street at the far corner and was gone. Taking
a deep breath, I turned and crossed the threshold back into the house.
My family remained where I had left them, except now they all pretended to be
relaxed. All except mother; she fixed me with her eyes the second I came into the
room. “He is much shorter than I expected,” she said. Such a statement about some-
thing physical was standard procedure. Mother had been doing this since my first
date. Except this was no date and this was no boy; it wasn’t even a man in the bio-
logical sense. My mother and I had shared everything when I was a girl. Neither she
nor I had ever mentioned this option. A little shudder went up and down my back.
“Cool,” said my little brother, Christopher. He looked at me with his big brown
eyes and smiled with approval. That helped. He loved robots and had a small
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
64
collection he kept in his room, so I wasn’t surprised by his enthusiasm. SamuelG
had asked to meet the robots, which delighted Christopher to no end. I don’t know
what they talked about for those few minutes I remained downstairs to help set the
table for dinner, but I do know that samuelG was serious about meeting any kind
of robot, AI or not. He and Christopher could probably have talked all night, or until
we called them to dinner.
“He doesn’t have much of an appetite,” said mother. She knew, of course, that
androids don’t need food; but in her world, much of the ability to find pleasure
comes at the end of a fork. I knew better than to pursue this line of commentary.
My father was apparently reading a book; one made of paper no less. Whether
he was actually reading it or not was difficult to tell. Probably not; at this point
he would be listening to the tone of my mother’s voice to determine where the
line of questioning was likely to go. Like a savvy lawyer, he knew when to let the
judge have her say. My mother often accused my father of condescension precisely
because he preferred to wait her out.
My elder sister, Helena, fidgeted with her shirt. This was not a good sign. There
were times when I could imagine her still tied to my mother by the umbilical cord.
When my mother was nervous, Helena would fidget with her shirt. The question
was, why was mother nervous? This may not be what you’re thinking. My mother
is not some Luddite bluestocking who hates the very idea of androids, much less
her daughter joining with one in a permanent relationship. Her concern would be,
I hoped, that in marrying an android I might be consigning myself to a life of weak
emotions and little passion.
My mother had many passions in her life; my father being just one of them.
When I say passion, I mean that in the full-throated way of a great Fado singer,
which she was. Fado is supposedly about sadness and loss, but in her Brazilian
fashion, my mother sang against the mournful strains with vigor and passion. Red
dresses, red roses, red lips and a lust for life was how she described it. It worked for
her; she assumed it would work for me.
“Can he dance?” My younger sister, Theresa, betrayed great doubt about this
in her voice. Like my mother, she could move like a cat and flash her eyes like they
were on fire when she danced.
“Yes, he can dance really well,” I replied. “I’m teaching him.” I was very thankful
our living room was too small for a demonstration, or they would have asked him
to dance right there. SamuelG could dance well enough but not with the sense of
feeling my sister and mother would want to see.
As usual, mother saw right through to the point. “I am sure he can dance.
Machines have been dancing for decades. The question is to what tune will he
dance in the future? Yours or his?” She held out her arms like a Flamenco dancer
and clapped her hands sharply three times. My heart sank. I could feel where this
was going.
“Can he sing, or play the guitar?” Theresa ran down the path my mother had
chosen. There was more than a hint of accusation in her voice, as she knew the
answer to her question and didn’t like it. She knew as well as I did that an AI singing
was only imitating notes and words from a recording that played somewhere in its
head. Often it was technically correct but without inspiration.
“Theresa, you know that androids do not have true vocal cords; they can’t
reproduce the timbre.They also can’t mimic the whiskey-soaked tones of the blues,
or control their simulated vibrato like a Metropolitan Opera soprano. The effect of
an AI android singing is most like people singing in a cold-water public shower.
“Then how will he communicate his love for you? If he can’t dance it with you
and he can’t sing it for you, what will he do?” Theresa can be terribly naïve for a
1010101010101010101010101010101010101010101010 65
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
girl of sixteen. Oh god, I thought, I can’t tell her what he does. I blushed. Of course,
mother saw that immediately and began laughing uproariously, which made me
blush even more.
Through her laughter mother said, “So, he’s not as dull as he seems, or are you
teaching him how to make love as well?”
This was too much; I felt like exploding. My father, bless him, suddenly said, “I
believe he is a poet.” It was brilliant; the life of my mother and younger sister was
set to perpetual music and little else; but they had to respect poetry—the words
their songs accompanied into the world. “Yes, he is a poet and he draws upon all
words in all languages in a way I have never heard before.” It was true too; samu-
elG was fascinated by human languages, which he loved to combine in ways that
sounded musical as well as meaningful. That thought gave me courage. All my life
my mother used her well-formed opinions to hold us all in check. It was like she
cast a spell, just as a great Fado singer casts a spell on her audience. Only now I
was no longer in thrall. I was ready for a brave new world my mother chose not to
experience.
I looked at my father, who smiled warmly and nodded. I looked at my little
brother and he gave me the thumbs-up. I looked at Helena; she had popped a but-
ton. Theresa looked at me more than a little astonished, not quite understanding
what just went through my mind. Then I looked at mother. She knew.
1.7 GENDER DIFFERENCES AND GENDER
DISCRIMINATION
“We are more comfortable thinking of our AI projects as the opposite sex or something
neutral. And that is reflected in the names we give them. That is why we see projects like
Alice and Leonardo. I wonder, who came up with Watson. I kinda hope it was a woman.”
—ANONYMOUS
It’s hardly a confession and certainly not unique to remark that for most people sex
is both attractive and difficult. I’m not talking difficult because of taboos and prud-
ish hesitation, that’s so 20th century. No, I’m talking about the modern world of sexual
relations where anything goes and where simply finding one’s own ground is a major
undertaking. Sexual relations were almost never easy, now it’s still difficult and become
increasingly more complicated.
By the way, we’re talking sex of and sex with, as in gender and sexual activity. Both
apply to dating and relationships of course, but the intriguing element for this chapter is
what does sex, in both senses, mean in a relationship with artificial intelligence?
AI AS GENDER BENDER
Let’s start with this: gender in AI, whether without a body (avatar) or corporeal
(android), is male, female or neuter. You probably noticed that one gender difference
from most human beings, neuter.
Of course, AI in any format have no native gender. There are no genetics involved;
only gender-related mental models and gender-specific physical properties, if any gender
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
66
1010101010101010101010101010101010101010101010 67
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
at all. In short, AI are developed or built to have a gender, or not. Many can change gen-
der on the fly, if necessary. They can be true gender benders.
Neuter AI, or more accurately AI without any gender characteristics, are more com-
mon than you might think. There are occupations such as financial trading or farming,
where having a gender is either irrelevant or possibly negative. There are also situations
in relationships between AI and people where the person involved might not want the
complications of gender, for example as a dispassionate judge. On the other hand, we’re
accustomed to human beings with identifiable gender. How would you react to an AI
that looks human but has no discernible characteristics of gender?
MEET GORT AND KLAATU
To frame an answer to the question of genderless AI, I’ll reach again into the trove of
science-fiction movies. This is like using shorthand for people familiar with the movie,
which I hope is true in this case. If not, hopefully the description is enough.
Consider Gort and Klaatu. Gort was the alien robot from the classic science-fiction
movie The Day the Earth Stood Still (the 1951 original, not the dreadful 2008 remake).
Gort was eight feet tall, stolid and spooky—goofy-looking, yet a robot with enough
power to destroy a planet. Gort was literally a galactic peace officer come to Earth to
make humans become less warlike. Gort was vaguely humanoid with a head, two arms
and two legs, but blocky and with non-descript features—obviously an asexual robot.
For comparison, Gort’s companion Klaatu, played by the Englishman Michael Rennie,
definitely looked like a human male, a rather sexy one at that. Lean, distinguished, reserved,
pepper-haired—the English make good-looking aliens. It was obvious that Mrs. Benson
(Patricia Neal), the female lead in the movie, was very attracted to Klaatu, although Klaatu
provided no gender-related emotional response to her. He simply looked and acted like a
good if somewhat odd man. The movie contrasts the attitude of the general public, which
sees Klaatu as frighteningly alien, with that of Mrs. Benson, who eventually understands
this is an alien, but one with characteristics that are quite human and even noble.
In Gort and Klaatu aliens take different forms. Neither actually has a gender; how-
ever, the one that looks male is quickly accepted as both human and male. Where does
that put AI that have no gender?
Most of us have known people, often very old people, for whom the notion of gen-
der doesn’t have much significance but there is no question of their humanity. We also
probably know people who are unusually androgynous, whose gender is not obvious or
possibly ambivalent. Again, there is no question of their humanity. Most people would
grant there are important human characteristics for which gender is essentially irrel-
evant. Compassion and empathy, for example, are associated with humanity but not
deeply associated with a specific gender. By exhibiting some of these characteristics, it is
possible for a robot or a projected AI image to be taken for human, with or without any
outward signs of gender. So the answer to the question of whether neuter AI can also be
taken for human is yes, as long as they look human and exhibit human characteristics.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
68
SO WHAT ABOUT SEX?
Neuter AI have their place. Does this place include sexual activity? Don’t jump to con-
clusions on this. Since I’ve been in a sort of Twenty Questions mode in this chapter,
here’s another question: Can sexual activity be performed without gender?
Of course it can: read a history of eunuchs or note that the sex toy industry thrives.
There are many aspects of sexual activity that don’t require the actual biological para-
phernalia or a specific gender. For example, neuter AI androids can have genitals with-
out having the accompanying gender characteristics. However, this is unusual. For most
people, sexual activity involves gender, as a matter of preference if nothing else. This also
applies to relationships with AI.
For people, gender is obviously the most important where conception is the point
of sexual activity. It’s different for AI. Female-gendered AI are not capable of biological
conception and gestation. However, as I’ll elaborate shortly, male-gendered AI might
perform artificial insemination. As I wrote above, AI can be gender benders.
A BRIEF HISTORY OF MECHANICAL SEX
By mechanical sex I don’t just mean sex with a mechanical device, I also mean in the sense
of sex performed mechanically—that is without emotion or much personal involvement.
A good example is masturbation. While almost everybody masturbates, it is almost by
definition performed mechanically. Sex with yourself requires only an effective technique
of stimulation and no more complex purpose than achieving orgasm. Of course, fantasy
is often part of masturbation, so there are times when personal involvement or even emo-
tional involvement may be active—but people mostly masturbate to get off, period.
So when I write ‘mechanical sex,’ I mean sex without love or probably any well-
focused emotion whatsoever. That kind of sex is as old as mankind, and probably older.
We know our contemporary close primate relatives masturbate at least as avidly as we
do; and presumably so did our paleological ancestors.
Part of masturbation is the use of tools…I guess I can use that word, as when it
comes to masturbation we use any appropriate tools that come to hand such as fruits,
vegetables, flasks, bladders and shaped objects of many kinds; hands and fingers are, of
course, some of our favorite tools. However, being the tool-using technological creatures
that we are, our use of mechanical means for masturbation more or less tracks our tech-
nological advance.
The first such tools are mentioned in Greek and Roman literature. There was little
or no innovation in sex tools for many centuries until the dawn of the industrial age
between 1750 and 1850. Suddenly, with the advent of mechanical clockwork devices
and eventually self-powered engines, the engineering genius of humanity started to pro-
duce mechanical sex devices. They were designed for both men and women and were
frequently condemned and made illegal. However, they were profitable to make, and so
innovation continued. By the end of the 19th century, the range of self-activated sexual
apparatuses made largely by and for Western countries was quite sophisticated and
diverse. Some equipment even used the new-fangled energy source known as electricity.
1010101010101010101010101010101010101010101010 69
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
I’ll note in passing that during the 19th century most of the sex tools for women (one
form of vibrator or another) were considered ‘for medicinal purposes only’ and were
generally available only through the services of a doctor. This was considered acceptable
practice even in the more puritanical countries such as the United States and England.
Doctors, almost entirely male, found the practice both lucrative and beneficial. However,
by the 1920s the main medicinal purpose, which was treatment of women’s ‘hysteria,
was no longer recognized as a valid medical problem. Doctors lost their access to a prof-
itable niche.
During the twentieth century, mechanical sex devices made the transition to elec-
tricity (which was often battery-powered) but did not fundamentally change much. Fol-
lowing the Second World War, the introduction of plastics and eventually electronics
created another boom in innovation similar to the mechanical boom of the 1800s. This
largely coincided with massive changes in public attitudes about sex and particularly
the role of sex in the lives of women. By the end of the century, the use of inexpensive
mechanical devices (mostly plastic sex toys such as vibrators and simulated vaginas) was
commonplace, if not widely acknowledged.
More elaborate sex devices, some commercial, some DIY (do-it-yourself, for doing
yourself) were readily available. This was also the time of the first appearances of robotic
sex devices. Crude and expensive were the two most appropriate adjectives for sex robots
(sexbots) but, as ever, novelty had its adherents. In this case, Japan led most of the devel-
opment, combining the Japanese ability to objectify sexual appetite with talents in elec-
tronics, manufacturing and especially robotics.
The sex robots of the late 20th and early 21st centuries were almost entirely brainless
(meaning little or no AI was applied) and their mechanics were mostly crude electronics.
Still they represented a bridge, or a framework, for the eventual incorporation of much
more sophisticated electronics. These would include the integration of AI-driven behav-
ior and physical structures using nanotechnology and synthetic biology.
Well into the 21st century, the ‘mechanical sex device’ began to lose its mechani-
cal flavor and became much more ‘lifelike.’ What was already acceptable technology for
many people in the 20th century became even more acceptable for a greater number of
people in the 21st.
HUMAN IS AS HUMAN DOES
The critical point is that at least in certain areas of relationships (sex being one of them),
behaving like a human has become more important than looking perfectly human—
and this includes gender. Behavior, however, is where things get tricky.
Maybe it was the dim lighting, or the three drinks, but I told him I thought he was
cute. He turned on his bar stool to look at me with an expression that said he was
trying to understand what I meant by cute. So I quickly added “handsome.
I know, these are loaded words; cute is for kittens and handsome is for hunks.
Either way, I’m giving signals that should be unmistakable. He looked like he got
the signal, but neither his mouth nor his body language said anything else. That
should have been my cue, but I’m the type who likes to make men react, one way
or another. No reaction becomes an instant challenge.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
70
“I get it, you’re from New England. They don’t unfreeze until mid-July. Or do I
need to send you email? We could meet again in say, three days.” Blink. Now he’s
not saying anything, and the bartender is standing there, tapping me on the arm.
“Sister, that one’s an android; and not a very advanced one at that.
I suspect most of us have had the experience of discovering someone attractive, try-
ing our best to flirt or perhaps something more aggressive only to discover that the per-
son is of an incompatible sexual persuasion. Gay, straight and occasionally transgender
mixups are not uncommon despite the fact that most people have their sexual antenna
tuned to the utmost in dating and hookup situations.
A guy walks into a bar and sees a beautiful female android sitting alone. He says,
“Can I buy you a drink?”
She replies, “Sure, but I don’t drink and anyway it won’t do you any good.
He orders drinks for himself and asks her, “Can I order something for you to
eat?”
She replies, “Sure, but I don’t eat and anyway it won’t do you any good.
All this is expected of androids so he downs his drinks, orders some dinner for
himself and chats with the beautiful android for nearly an hour. Then he asks her,
“Why don’t you join me at my apartment?”
She replies, “Sure, anyway it won’t do you any good.
Now she’s really got his interest piqued, so when they get to his apartment he
almost immediately says, “You are the most beautiful female android I have ever
met. I think I want you for my wife!”
She replies, “Ohhhh, that’s different! When is she coming home?”
Mistaken first impressions are part of the scene; is that also true for AI? Mostly, no.
AI in all forms, avatar or android are usually constructed to present a specific gender.
That gender might be changeable, but then the change is obvious. What isn’t so obvious
is that AI are also constructed with complex mental modules that determine personal-
ity and a host of other characteristics—including sexual preference. This is no more
obvious on visual cues than it is for people: it can be detected by “gaydar” or whatever
observational powers are available—but not always.
BEYOND FIRST IMPRESSIONS
First impressions like the first date are important, but as almost everyone will tell you,
it’s the later experiences that are usually more revealing. This is true in human relation-
ships, certainly when it comes to as we say, ‘finding out what somebody is really like.’ It’s
such a common refrain that people put their best selves forward during early relation-
ships; but you have to see them in the home setting, not dressed up; without make-
up; under-slept; stressed by work and concerned about their own problems, before you
really know (anything) about their true self.
Relationships with AI aren’t like that, mainly because AI are not putting up a front.
In terms of character, what you see is usually what you get. If there is a parallel with
getting to know AI as you would get to know a person, it’s that the depth of an AI’s per-
sonality (I’m tempted to say humanity) varies. Also what an AI reveals to people varies.
1010101010101010101010101010101010101010101010 71
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
Obviously they can’t and don’t tell you everything—there isn’t time. AI are not neces-
sarily skilled in understanding what and how much they should reveal to their human
partners. They make mistakes, as do we all.
What is generally agreed is that when it comes to sex—gender and sexuality—AI
are both gifted with a great deal of knowledge and ability and cursed with an under-
developed sense of what is appropriate. This can lead to some odd moments:
Let me tell you right away this was not my first experience with an android. I like
sex with androids and have had lots of experience. If there is one thing I’ve learned,
don’t expect the android to lead. It’s not that they can’t but I haven’t met one yet that
was inclined to take the sexual initiative. That is, until I met fredZ.
As you certainly know, AI are nothing if not learning machines. Learning is what
they do whenever not specifically doing something else, and in any case, they con-
tinue learning in the background no matter what they’re doing. Ever had that feeling
that even in the heat of action your partner wasn’t quite ‘there?’ That’s pretty much
standard for AI. That’s been my experience, except for fredZ.
I don’t know if the people who built fredZ had a special contract, or fredZ had
some very unusual early experiences, but his mental patterns are…different. No.
They’re weird. He thinks he is a superman of sex, which of course he is. All androids
are, compared to their human counterparts. They never tire. Their reservoir never
runs dry. They don’t care if something is old or new in terms of technique. They don’t
have moods. They don’t need to get up and go pee. They’re not thinking of doing
something else with you in the back of their mind. So yes, from at least the physical
perspective they are supermen; but fredZ insisted that he take me to the next level.
Somewhere, somehow fredZ came to believe that I was capable of more
intense, creative and satisfying sex. I can’t place the day but it was weeks ago,
shortly after we met, that he began what I can only call a campaign to up my game.
At first I had no idea what he was doing. This wasn’t a matter of macho demands
and male posturing—not at all. It was more along the lines of, “Come on, you can do
it one more time”—like a personal trainer telling you to do one more rep. Always pos-
itive, always coaxing nicely but very firmly that I could make this move or that move
and the effect would be better. It usually was better, except that I was wearing out.
So here is this lovely, well-endowed android who is trying to get me to the
ceiling in my responses, with apparently the best of intentions—and I’m dying, flat
out wasted. I’m wondering if I’m losing it, my magnificent sexual urges, or if in fact,
fredZ really is superman and I’m failing to fly along.
Frankly I am ‘that kind of girl;’ I don’t shy away from any kind of sex and I’ve got
pretty good stamina. But I’m also aware there’s more to sex than physical activity. It
is not the Olympics. Or if it is, it’s a team sport. In fact, it’s like making music. I know
from experience that the best sex is a duet. For me it’s a matter of rhythm and that
rhythm must be shared; rubato in synch, if you know what I mean.
The problem is fredZ ain’t got no rhythm. He can count. That’s how he seems
to know I can always do more; but counting isn’t rhythm. I try to tell him that. He
knows what rhythm means. He can even tap out a beat, but attach it to feelings?
”FredZ! Stop counting and get with the rhythm! Can’t you feel it?” No he couldn’t
really; but he wanted to learn (of course). So now I’m leading and our sex is awful.
It proves that when it comes to AI, we’ve all got a lot to learn.
I hope the idea is coming across that relationships with AI can start out feeling
quite human, but don’t always stay that way. Gender for AI is something they put on,
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
72
almost like clothes. Because they are designed to mimic human behavior and in general
are made for human tastes, AI have many similarities with human beings—more than
enough to endear themselves to many people. But there are differences and they are
significant, sometimes in ways you wouldn’t expect.
SEX WITHOUT OFFSPRING—WITH ONE EXCEPTION
As I mentioned, an obvious difference between AI and humans, at least in most people’s
minds, is that AI can’t have babies—at least not yet. From that it is assumed that children
and parenting are not part of either the gender or sexual activity that goes with AI. That’s
not actually correct.
Of course, whatever the gender of an AI, one thing is certain—with an important
exception—sexual activity does not lead to pregnancy. Obviously, AI in any form—ava-
tar, android or whatever—cannot produce children. However, since some female human
partners can have children there is one important exception to the rule of “sex but no
children.” Some male AI partners are capable of artificial insemination.
As a matter of technology, this has been possible for decades. Over the years the
mechanics of robotic delivery have become increasingly realistic via simulated male
orgasm and ejaculation, and the effectiveness of the artificial sperm-carrying environ-
ment improved. Many women can and do become pregnant through this method and it
is having profound personal, social and cultural effects.
With the advent of in-vitro fertilization (IVF) in 1978, the process of fertilization by
means other than the traditional male-to-female insemination not only became techno-
logically feasible, but with a rapidity that surprised many people, was generally accepted
and widely used. When it later became possible for male androids to carry live sperm
and inseminate women through a simulation of the traditional method—it was almost
retro. Not that there weren’t objections, of course, but the acceptance of the practice was
also rapid. As a matter of technique, it was less ‘mechanical’ than surgical implantation
and definitely more enjoyable for women.
More significantly, it provided a compelling reason for some women to accept
android partners. Not only could a woman choose which sperm to use from extensive
sperm bank catalogs, but she could also have it delivered in a more personal way, poten-
tially by a committed partner. Psychologically it was liberating, even for women who
didn’t participate. It took several decades to unfold, but many believe it was a crucial
element in changing the balance of many societal norms in favor of women.
The corresponding decline of the male role in conception, however minimal in real-
ity, was eventually perceived as a threat, which led to seemingly endless legislative and
legal battles in many parts of the world. Very often the objections were couched in reli-
gious terms; however, many religious groups were conflicted as well. Most notably, the
Catholic church was internally divided on the question of android insemination, with
those who favored the traditional role of the human male (patriarchists) squared off
against those who believed that technique was irrelevant to a Catholic conception (spir-
itists). The controversy became especially passionate over the re-opened issue of Mary’s
immaculate conception of Christ.
1010101010101010101010101010101010101010101010 73
ARE YOU READY TO FALL IN LOVE WITH A MACHINE?
For AI and android robotics, or more accurately, for those who designed and built
them, the seemingly unspectacular technical ability to deliver active sperm changed
some of the dynamics of the industry. For one thing, it made gender, especially the male
gender, more important to the business prospects. There was increased demand for male
androids, for women to become pregnant of course, but a premium developed for AI
avatars or androids that could also be successful companions for women. I should add
up to that point there was the tendency (bias) in the industry to provide ever more
sophisticated female AI robotics that catered to men in the usual tradition of pornogra-
phy. Ironically, this demand prompted increased resources for gender-oriented research.
It resulted in much-improved AI conceptual gender models for the technology in both
male and female avatars and androids. It wasn’t the first or the last time that sex moti-
vated technical development.
NOT THE END OF GENDER DISCRIMINATION,
BUT THE BEGINNING OF THE END
The choice of using a male android for insemination represented something of a break-
through for having relationships with AI. What this choice made obvious was that now
women or men could choose a partner of any gender or no gender at all, and could
possibly have a meaningful relationship that might include children. It gave women a
powerful new option that changed the balance of natural selection. If it did not signal
the end of gender discrimination, it at least marked another point where discrimination
based on gender was losing its meaning. Now there were more viable alternatives.
I don’t need to emphasize that discrimination based on gender is very old and
persistent. Woman has been the object of discrimination for millennia. The whys and
wherefores are legion. What’s most relevant in this guide is that relationships with AI
are affecting gender discrimination of all kinds, but thanks to artificial insemination by
male androids and avatars, the greatest effect is on the discrimination against women.
It is said, and it could be true, that the greatest contribution of AI to decrease gender
discrimination is that they don’t discriminate. There’s another part to that: AI in any of
the forms (avatar or android), have a gender mostly in response to human wishes. They
are not committed to the notion of gender for reasons of biology or genetic inheri-
tance. In fact, quite a few AI have no gender at all and this makes little difference to
them individually. It makes sense that in any society where there are a large number
of AI in personal relationships with human beings, the culture will be affected by their
non-discriminatory behavior and by their rather indifferent attitudes toward gender. In
this new environment, we find not only new personal options, but new social, legal and
economic allies.
1010101010101010101010101010101010101010101010
1010101010101010101010101010101010101010101010 75
You are ready, now what?
“Be careful the environment you choose for it will shape you;
be careful the friends you choose for you will become like them.”
— W. CLEMENT STONE
YOUR REASONS FOR DATING AI AND PERHAPS TAKING IT MUCH FURTHER
into a relationship are your own. That doesn’t make your reasoning unique—which is
why this guide exists—but ultimately you’ve chosen to explore a relationship with AI
and you’ll do so on your own grounds. From here on I’ll write as if you’ve come to the
conclusion that you’d like to at least dip your toes into the AI experience. I’ll do my best
in this section to describe and explain some of the things you can do to prepare yourself
for that experience.
I don’t want to give the impression that getting ready for an AI experience is so very
different from dating the girl or guy next door. There are similarities, and a lot of people
also prepare themselves for a date with the girl or guy next door. However, there is no
question that dating AI is more exotic than dating a neighbor. Developing a relationship
with an AI is not the same as “shacking-up” with another person. Obviously, AI aren’t
human and there are, as you might guess, other important differences. There’ll be much
more about that in section three. For now, let’s concentrate on you.
2.1 PREPARING YOURSELF FOR THE UNEXPECTED
“Amici, diem perdidi”
—VICTORY LOVES PREPARATION, A LATIN PROVERB
Is a personal relationship with AI going to produce the unexpected? You expect it will,
right? That could be one of the reasons you’re interested in AI in the first place. It’s not
that relationships with other people don’t have their unexpected moments; but AI, as far
SECTION
2
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
76
as most people are concerned, are like another species and present an unexplored terri-
tory of mind and experience. Of course there will be surprises.
What kind of surprises, do you suppose? If you could speculate accurately, they
wouldn’t be surprises; but a bit of thought about it might prepare you for the future.
What I mean goes something like this:
Whether you’ve read the book or seen the movie, this line from Forrest Gump is
probably familiar:
‘My momma always said, “Life is like a box of chocolates. You never know what
you’re gonna get.”’
Life is full of surprises in a simple candy metaphor; but rather than leave it at that,
let’s take apart this statement. For example, if life is picking chocolates, it’s certainly a
pleasant outlook on life. Is life like that? Bittersweet, maybe. In any case, the choice is not
totally random because you know the box contains chocolates, not other kinds of candy.
In fact, if you’re like most people who are familiar with boxed chocolates, you have
favorites. You can tell the nut ones from the creams and so on. Some people even know
that chocolates are sometimes coded, possibly with an “O” on top of the orange-filled
chocolates. So only some of your choices are surprises.
1010101010101010101010101010101010101010101010 77
YOU ARE READY, NOW WHAT?
The point is, as much as we say we like surprises, we like being in control of our
choices even more. This also goes for our choice in partners, which may include AI.
A relationship with AI is like a box of chocolates: you may be knowledgeable about
what you’re getting, but be prepared for the unexpected.
Doesn’t have as great a ring to it as the original, but then, it’s a modern decon-
structed metaphor.
WHAT DO YOU MEAN, ‘UNEXPECTED’?
When I say to prepare yourself for the unexpected, I don’t mean it’s something like the
boogeyman jumping out of the shadows, or even that an AI is unpredictable. It’s more
sophisticated than that, something along the lines of that old joke about Mrs. Webster
catching her husband with the maid: “Mr. Webster, I’m surprised!” His reply: “No my
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
78
dear, I’m surprised; you’re astonished.” The unexpected with an AI might be nothing
more than a different moment, perhaps an odd perspective or a desire to do something
that seems out of character.
When preparing yourself for AI, you should take into account that they aren’t exactly
human, so their responses and behavior will sometimes surprise you. It also should
include the general tendency of AI not to do the same thing twice. This requires a little
explaining.
As you probably know, AI have very good memories and more to the point, these are
permanent memories. We’d say they have photographic memories, but in this context
I’m not sure what that means. AI don’t record every visual detail of every moment of
every day, any more than people do. Nor do AI remember every fact they’ve ever heard,
read or absorbed through transmission. But when it comes to experiences, to paraphrase
Dr. Seuss: AI are like elephants, they remember one hundred percent.
Since they never really forget past behavior and they’re predisposed to be con-
stantly learning, AI usually attempt to not repeat behavior. This is not to be overem-
phasized at the trivial level. AI follow patterns of behavior much like people. They say
“hello” when they meet, shake hands, smile and so forth just as often as we do. How-
ever, an AI knows when it is repeating and if the situation is convenient or appropri-
ate, they’ll try something new—although sometimes ‘new’ is so subtle that only the AI
knows it is new or unusual. Nevertheless, it’s the constant desire to try something new
that is one of the outstanding characteristics of most AI. This provides a solid motiva-
tion for people to prepare themselves for the unexpected. If you’re going to have an AI
for a partner, you’d better be able to match (or at least tolerate) their constant learning
and innovation.
FOR BETTER OR FOR WORSE…
Life being what it is, not all unexpected events or surprises are good, much less enjoy-
able. Personal relationships have their ups and downs along with the usual slings and
arrows that fortune throws your way. This is reflected in a protocol familiar to most
human marriages, called ‘the vows.
Most wedding vows are optimistic. They’d better be. However, most wedding vows
also remind the bride and groom of the binding nature and some specific conditions of
the relationship. In the most common of Anglo-American vows, the phrase “for better
or for worse, in sickness and in health” is among the reminders that marriages are not
all roses and romance—life, good and bad, intrudes. The vows make it clear that the
marriage relationship needs to persist through good times and bad times, which is easier
vowed than done.
What about relationships with AI: are they also for better or for worse, in sickness
and in health? If so, what does better, worse, sickness and health mean in the context of
AI? In a way this is a proxy question for whether there is a darker side to a relationship
with AI. With human beings there is always the unexpected occurrence, for example,
cancer or an accident with the potential for death. AI don’t get ‘sick’ but they do break
1010101010101010101010101010101010101010101010 79
YOU ARE READY, NOW WHAT?
down. Androids can be run over by a bus. AI computer sources can become infected by
malignant viruses, which is as close an analogy to human illness as one needs. So yes,
there can be unexpected bad events, even with AI.
As it is with human relationships, it’s not healthy to dwell on potentially bad things.
It’s better to be realistic. AI may not be biological but they also suffer the slings and
arrows of outrageous fortune, though perhaps not as often, or as easily, as those of us
on the mortal coil.
WHAT DO YOU EXPECT?
Thus far I’ve been describing AI doing something unexpected, which they can and will
do. There’s another side to this, which is at least as important—What do you expect?
Tonya Hurley, best-selling author for young adults: “If you expect nothing, you can
never be disappointed. Apart from a few starry-eyed poets or monks living on a moun-
taintop somewhere, however, we all have expectations. We not only have them, we need
them. They fuel our dreams, our hopes, and our lives like some super-caffeinated energy
drink.” [from Homecoming]
Whether AI or human, starting a new relationship doesn’t begin with a blank slate, a
tabula rasa; you have expectations. Sometimes these expectations are specific, for exam-
ple you expect your AI date to know about current events. More often, expectations are
quite generic. You expect an AI to be intelligent. When somebody sets you up with a
blind date, you still expect the other person to be of a certain gender and most likely also
of a certain sexual persuasion. The same applies to a date with an AI.
These are expectations that I believe are pretty much universal. No doubt you also
expect your date isn’t a homicidal maniac or some other kind of sociopath. You never
know what might actually show up, but it’s reasonable not to worry about it. What I
mean is that you have realistic expectations. You can realistically expect that your date
isn’t a sociopath or for that matter a billionaire. I suppose if you were going to date a bil-
lionaire, you’d know about it—wouldn’t you? Knowing ahead of time about a sociopath
is a bit more difficult, since that’s not likely to show up in the profile. Or are you one
who only likes to find dates by looking the potential person in the eyes (like in a bar),
or by using online services that provide comments from people who have already dated
them? The main thing is that what you expect often depends on whether you’ve done
any homework or research. Some people don’t like doing personal research. For them it
stifles spontaneity; they’d rather be surprised. Perhaps fearing the random sociopath or
maybe just out of curiosity, other people like to know as much as they can about a pos-
sible date. Is one of these approaches a better strategy for dating AI?
“He wasn’t at all what I expected.” Marge is usually sympathetic, but she’s heard me
say this maybe a dozen times. She probably expected it. “The last dozen guys you
went out with weren’t what you expected,” said Marge, “and by now you’ve got to
be wondering if you’re expecting too much.” I told her, “I don’t know about that. I
certainly didn’t expect him to not notice my neckline.” Marge looked at me, noting
again the plunging V of the evening dress: “I warned you about that dress on the
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
80
first date.” She had warned me, but this time the effect was negligible; not what
either of us expected, I think.
I know dates in the Cloud are different. Even the best of holographic recorders
don’t do justice to people. But still…
Marge must’ve been tracking my thoughts, because she said, “After all, he was
an AI avatar. I mean, they’re not known for drooling. A lot of them don’t show much
interest in sex at all, at least not on the first date. It’s not their priority.
I didn’t want to admit it was my priority, or one of my priorities. Marge was
experienced in these matters and she would probably laugh. She knew a lot better
than I did just how complicated it was to have a physical relationship with an avatar.
Complicated, but not impossible. Maybe I did expect too much.
“I still don’t understand why you didn’t use your own avatar,” Marge said this
with just a touch of sharpness. I guess her own experiences made her cautious.
“Marge, I just wanted to get an honest impression. You told me that AI are more
interested in the real deal. So I figured seeing me in the flesh, kind of, would get a
better response.
Marge sighed. “I bet you looked at his pictures and barely scanned the profile.
Did you even remember what his preferences were?”
Actually, I sort of remembered, but I expected the preferences wouldn’t be very
accurate. Most people don’t put down what they think they can realistically get;
they usually go for some kind of ideal. “I didn’t expect his preferences were realis-
tic, Marge.
“Sally, this was an AI. They don’t make things up; or at least not usually. Mak-
ing things up is what humans do, right? I make things up on profiles all the time,
it’s part of the game—except when I want to date an AI. They check things and they
study you; it only takes them an instant. … I’ve told you all this before. So how did
the rest of the date go? Was he interesting?”
“I don’t know. Really. All right, I guess. I liked his face in the pictures; but his live
avatar was…different, somehow.
“Not what you expected,” said Marge. “I’ll bet you thought he had a kind face,
but watching the avatar—well, there wasn’t much empathy there at all.” I must’ve
looked kind of shocked because Marge immediately said, “It’s happened to me,
more than once. I quit expecting the pictures to reveal much of anything about an
AI’s avatar. When you interact with them, it’s almost always different.
Marge was sure right about that. My one and so far only experience left me
wondering just how different AI really are. I know they’re designed to interact well
with people…but I have to admit I’m not very ready to interact with AI. I mean, I
don’t know much about them. I don’t really know what they might be looking for
in a date. That’s one of the things about dating AI: I think they don’t have a list of
expectations, or if they do, it’s not a very important part of the experience for them. I
guess they’re most likely configured to take people as they find them. You just have
to expect an AI to be different like that.
Most modern AI have personalities, which means they have some preferences, opin-
ions and personal feelings. Regardless of their mental models, most AI at least give the
impression of making choices. In terms of a relationship this usually adds up to AI say-
ing or doing things you don’t expect. Mostly this is just normal variation, much like you
would expect from a human partner. At times though, AI behavior will seem profoundly
different, and for some people this can be unsettling. Fortunately there has yet to be
1010101010101010101010101010101010101010101010 81
YOU ARE READY, NOW WHAT?
a case of a truly psychopathic or sociopathic AI, at least none involving fatality. Seri-
ously, that’s an important distinction in the favor of AI. It means that while you can’t
know everything about an individual AI any more than you can know everything about
another human being—with AI it is at least in one respect safer.
Does that mean the best strategy for dating AI is to take it as it comes—be sponta-
neous? The corollary is not to spend much time on personal research; since you don’t
need to know much about an individual AI and it might even get in the way of being
spontaneous. Really?
I wish there were some reliable studies on this, something to point to as indicating
how much you can trust that an AI will be ‘safe.’ They don’t have criminal records, a
history of violence, or even bad habits like alcoholism or drug abuse—that much we
do know. But there is only spotty anecdotal evidence that relationships with AI do or
do not affect people positively or negatively. Certainly AI behave differently and often
unexpectedly, and for some people that may be a problem; other people may consider
that really great.
Unless you overdo it, if your personal preference is to know what you can about an
AI beforehand—no problem. Likewise, not knowing much about an AI before the date
isn’t a formula for disaster. I will say though that AI will always know as much as pos-
sible about you.
WHAT DO YOU KNOW ABOUT AI?
While it may be a matter of personal preference to know a lot about an individual AI,
I think it’s different when it comes to knowing about AI in general. Part of what you
expect from dating AI: avatar or android, depends on what you know about AI.
• Doyouknowthehistoryofarticialintelligence?
• DoyouknowhowAIareconstructed?
• Whataretheprinciplesbehindanarticiallyintelligentmind?
• Whatarethecurrentstrengthsandlimitations?
• WhatarethedifferencesbetweenanAIavatarandanAIandroid?
I can think of at least a dozen questions like these, most of them pretty basic. You
don’t need detailed and encyclopedic answers, but having reasonable background
knowledge of AI is like knowing something about human physiology and psychology. It
gives you a framework to think about how an individual AI behaves, what the motiva-
tions might be, and how you might deal with the unexpected things that show up from
time to time. It can also help you develop realistic expectations for a relationship.
It figures that since you’re reading this book, you are a person who is already com-
mitted to knowing more. You are also a pioneer, the first generation of human beings to
have relationships with non-human intelligence. In some very profound ways, this is a
new experience and of course unexpected things will happen. The unexpected should be
considered part of the challenge and part of the fun.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
82
2.2 GET TO KNOW YOURSELF AND BECOME
A BETTER PERSON
“Know thyself”
—ANCIENT GREEK APHORISM ATTRIBUTED TO THALES
“It is not the Bible, upbringing or ethics that prevent me from doing evil, it is the prospects
of having my mind read by either another human, or worse… an intelligent machine. It will
happen in the very near future. In fact, it may be happening right now.”
—ANONYMOUS
Only those who are legends in their own mind cannot admit they could be better peo-
ple. Realistically, who couldn’t use a little—or a lot—of improvement? Getting to know
1010101010101010101010101010101010101010101010 83
YOU ARE READY, NOW WHAT?
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
84
yourself and becoming a better person ought to be a no-brainer. Except it isn’t a no-
brainer, it is very much a mental thing. Mentally we’re so very good at rationalizing.
For example, we can take a completely obvious personal insight such as ‘I need to spend
more time with my friends,’ and mentally postpone the thought with, ‘…but not this
month, there are just too many meetings and I’ve got to finish the golf lessons. Next
month though…’
Improving yourself, which is not the same as self-improvement, is a lot harder than
it should be. We have difficulty accurately identifying specific things about ourselves that
could be improved. Then, supposing we can find something to improve, we often fail to
find a way to do something about it. Worse yet, even if we identify a problem and figure
out a way to deal with it; we often lack the time and discipline to actually do something.
If I want to be really pessimistic, I could add that sometimes even when we do some-
thing, it doesn’t work.
Why is this important in the context of a relationship with AI? The answer isn’t
so obvious. Certainly in order to have a good relationship with an AI you can prepare
yourself ahead of time and you can work on things during the relationship. Improv-
ing yourself will help you attract and hold a relationship with AI. This is a truism for
any relationship, including a relationship with other people. As obviously beneficial as
improving yourself may be; most people won’t actually do it. That begs the question,
why should we make an effort for a relationship with AI? The answer to that question is
what this chapter is about.
KNOW THYSELF
Even though you could hardly function without a well-developed sense of self, most
people will admit their self-knowledge is imperfect. Why is that?
Humanity was probably puzzling about the self back when we first realized the thing
we saw reflected in a puddle of water was our own face. It was a surprise; such aware-
ness is not standard equipment. We are one of the few creatures on the planet that can
know itself. Self-awareness or just plain consciousness is not widespread among the liv-
ing things on Earth. It’s more widespread than we used to think, as scientific evidence is
accumulating that we are not the only self-conscious entity, for example, some primates
(apes), cetaceans (dolphins, whales) and elephants also have various kinds of conscious-
ness and self-awareness. Our self-awareness is still evolving, and that’s indicative of why
we’re not all that good at it. Yet we Homo sapiens can hold a bit of pride in that we’ve
evolved the most sensitive and wide-ranging self-awareness, assuming, of course, that is
a good thing.
I’ll accept that assumption; in fact, I’d accept the notion that self-awareness is a very
good thing. There is long-standing precedent for that idea: The ancient Greeks were
quite interested in the idea of the self and our ability to think about who we are as
individuals. The maxim “Know Thyself” (Gnothi Seauton) was used by many Greek phi-
losophers. What we often hear today is attributed to Socrates: “The unexamined life is
not worth living.” I suppose that could be interpreted to mean that if you don’t think
about who you are and how you live, then you might as well jump off a cliff. Ironically,
1010101010101010101010101010101010101010101010 85
YOU ARE READY, NOW WHAT?
as reported by Plato in the Apologia, Socrates said this as he was about to commit suicide
under threat from the state of Athens.
What is the big deal about knowing yourself? It’s a fair question, although the
answers are often intuitive. Babies don’t ponder the question, but they are perplexed by
having fingers and toes, which they learn are part of their body and are mostly under
their control. Infants must also resolve the very confusing existence of “others.” What are
these big creatures that also have fingers and toes, but make strange noises? They are not
“me,” but what are they?
And so it goes; to one extent or another we have to know ourselves, if only to figure
out what so many other things around us are. That figuring goes on for the rest of our
lives, whether we know it or not. What the philosophers think and most people accept is
that the better you get at self-awareness, the better you will be able to deal with the world
around you. Notice I wrote “deal with,” which at its most primitive level means survival.
Even today we speak of a person with no self-awareness as a disaster waiting to happen.
Then again, the interesting thing is not really people with no self-awareness, but the
various degrees of awareness that we have or that we can attain. That points back to the
question above: Why is our self-knowledge imperfect?
BE CAREFUL WHAT YOU WANT TO KNOW ABOUT YOURSELF
For one thing, not all people want to know themselves very well. It could be because of
deep, dark secrets. It could be fear of things unknown. While philosophers both ancient
and modern extol the virtues of examining one’s life and one’s self, literature, as it reflects
life, is rife with examples of knowing too much. A good example, sticking with ancient
Greek references, is the story of Oedipus the King. Although he was warned, Oedipus,
the new king of Thebes, persists in finding truth about the murder of the previous king;
until he discovers that through his own arrogance and impulsiveness, it was he that
killed his father, the king, and married his mother. Fortunately, gaining additional self-
awareness isn’t usually so grim.
The search for self-knowledge and the desire for self-improvement are often the
targets of humor. It figures. Humor is based on imperfection, and what is more imper-
fect than self-knowledge or our ability to improve? Here are a few old chestnuts that
illustrate just this:
“I tried to be patient, but it just took too long.
“I’m not conceited. I’m just unbelievably good.
“I’m a unique individual, just like everyone else.
“I never make predictions. I never have and I never will.
“With enough preparation I can be quite spontaneous.
As those verbal jokes above imply, the ability for people to self-deceive is at least as
great as their ability to learn true things about themselves. By their very nature, self-
deceptions are difficult to recognize and even more difficult to remove. And mostly,
people don’t want to remove them.
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
86
All in all, it’s difficult to really know yourself and especially to look at those things
about yourself that keep you from being a better person. It’s no wonder that most of
the world’s major religions consider self-awareness—enlightenment—and dealing with
personal flaws (sins or whatever) a big part of the religious experience.
SELF-AWARENESS IS ALSO ABOUT GOOD THINGS
So far, I’ve described how difficult it is to know yourself; mainly because most people
have aspects they’d rather not explore. But that’s clearly not the full story. Most people
have good things about themselves and part of the task of personal enlightenment is to
identify the good things and augment them, as possible, and add to their number.
“When it comes right down to it, all you have is your self. Your self is a sun with a
thousand rays.” —Pablo Picasso
There are thousands of adages, maxims, sayings and longer expressions about the
importance of improving yourself. These days, self-improvement is an industry. You
can approach self-improvement technically, like a face-lift; or spiritually perhaps in a
religious retreat; or haphazardly as most people tend to do.
“Thoughts become words. Words become actions. Actions become habits. Habits
become character. And character becomes your destiny.”
THE MIRROR TO HUMANITY
Much of self-awareness and self-improvement sounds like introspection. Certainly,
introspective techniques such as meditation can be very helpful (there is much more on
such techniques in chapter 2.5: Some strategies for developing an agile mind). There is
no question that a quest for awareness and improvement begins and ends with your own
motivation and abilities. But that too is only part of the story, because it is often (I’m
tempted to say always) other people who prompt you to look into yourself and to make
improvements. Sometimes this is active prompting, as the friend who sits you down and
says you are becoming too involved with an unhealthy lifestyle; sometimes it is passive,
as you realize that your friends are leaving you.
It’s often said, and we probably should take it more seriously, that human beings are
social animals. We evolved in groups and our psyche is developed through our interac-
tion with others. Modern life is something of an anomaly in that we have developed the
possibility for a large number of human beings to live more or less alone. It is a function
of our technology and material wealth that survival no longer requires an individual
to participate in a group. But it’s still unusual. Most people want to be part of a group,
if only a closely-knit family. In fact, most people, through work, family and voluntary
activity, participate in many kinds of groups—they are almost constantly socializing.
More often than not it is through socializing—interacting with others—that we
come to realize the need for more self-awareness. It is from others that we often take the
most influential cues about what to improve. We have our heroes, our models—people
1010101010101010101010101010101010101010101010 87
YOU ARE READY, NOW WHAT?
we admire and wish to emulate. We have those whose standards make our own failings
more obvious. In short, much of self-awareness and improvement isn’t a result of intro-
spection so much as it is the effect of your experiences with other people.
One of the most powerful motivations for self-awareness and improvement has
always been the desire for sexual, romantic and personal relations. How many times
have you considered who you are in order to find someone who is ‘right’ for you? How
many times have you done things like start a program of physical exercise so that you
can be more attractive? The effect of another person, even an unknown and unspecified
other person, can be a powerful motivation.
And that brings me to relationships with AI, the new mirror to humanity.
WHAT WE CAN LEARN FROM AI
Some may see it as ironic, but perhaps we can learn much about ourselves from one of
our greatest creations—artificial intelligence. While in many ways AI reflects human
intelligence and was developed to serve human interests, the end result is not exactly
human. Among the many different kinds of AI and AI robotics, there are sometimes
obvious and sometimes subtle differences from humanity. The differences increase as
AI learns and new perspectives are born. Thankfully AI can communicate with us about
those differences and we can learn from those new perspectives.
There are many things we can learn from AI to understand ourselves better and to
help us find ways of improving. This is true on both the level of human society and the
level of the individual. I’ve picked out three things where the presence of AI and robotics
is changing society and personal relationships. These are by no means the only things
(there’s a lot more in section three), but they are representative.
AI AND WORK ETHIC
Whether you know it or not, every individual develops a work ethic. I suppose it’s kind
of a continuum from ‘work as little as possible’ to ‘work as much as much as possible.
We all know people who love to work and are restless without something to do. We
also know people who are lazy and will expend effort to avoid work. Most people are
somewhere in between, and it varies over time. Attitudes toward work fluctuate with
health, age and personal finances (among other things); but most people do have an
overall approach to work that can be characterized by some mixture of avoidance and
enthusiasm—their personal work ethic.
Now, contrast human work ethic with AI.
The reason there are ever more AI and robotics in the world is because they have
economic value. They are developed because they make money for their developers.
Those who buy them from the manufacturer expect them to do work that returns a
profit. The cost of AI and robotics, which is very high, is a limiting factor to the growth
of their use. However, not only are their numbers increasing, but what they do and how
they do it is affecting the way people all over the world live. It is often said that we are in
DATING AI
10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
88
competition with AI and robotics, which in one sense is obvious. There are those who
say we should seek to learn from AI so that we can more effectively compete with them.
Still, if you think about it, what could we possibly learn about a work ethic from AI
that never get tired, never sleep and never stop working? It would seem that while an AI’s
work ethic might be enviable or praiseworthy, it is hardly a model for humanity. It’s not
something we can successfully emulate nor compete against. Besides, isn’t it one of the
reasons that AI make sense economically that they can do jobs requiring 24/7 attention
and people can’t?
Exactly. AI and robotics are forcing a fundamental re-evaluation of how human
beings work: How should we work given that AI and robotics can do many of the jobs
humans once did?
As has been understood for decades, robotics (and AI) are best at tasks that are
repetitive and do not require significant judgment or creativity. Robots started taking
over the assembly line way back in the second half of the 20th century. AI started replac-
ing certain kinds of clerical and service jobs in the beginning of the 21st century. Slowly
and not all that perceptibly, the kinds of work available to people have shifted. Certain
kinds of physically repetitive jobs have disappeared from the list, as have many so-called
white collar jobs such as auditing, proofing and researching.
IT’S MORE THAN REPLACING PEOPLE AT CERTAIN JOBS
Do AI get tired? In the human sense, no; however it has been discovered that AI ben-
efit from shifting context. Let me explain as briefly as I can: Intelligence and especially
creative intelligence is difficult to achieve and it’s become obvious that also in the realm
of AI, repetition and sameness are the enemies of most kinds of intelligence. While AI
are ‘always on’ and theoretically can work at one thing forever, they mostly don’t. They
switch their mental context; sometimes they focus on a task, sometimes they communi-
cate with one another, sometimes they have personal relationships with people—and so
forth. AI switch their focus, the context, to keep their intelligence active.
Human beings have to do something like this. Our intelligence, as that of almost all
biological forms of intelligence, needs rest and sleep. The brain doesn’t take a holiday, of
course, but it does relax, shut down some areas for a while and do some housekeeping.
We also prefer to vary our mental activity, often at very short intervals (especially for
those with attention-deficit problems). I mention all this because one of the best side
effects of AI development has been the greater attention paid to how human intelli-
gence works. While we’ve known for a long time that people require rest, play, variation,
change and social stimulation, our relationship with AI has helped a more nuanced and
hopefully beneficial understanding of what our intelligence requires to do its best—
especially in relation to our work.
What’s happening now is that our work ethic is adapting to the work ethic of AI.
Some of this is simply to avoid pointless competition. Some of it is to improve human
productivity. Much of it is simply a natural accommodation to a partner that also works.
If you are pursuing a relationship with AI, then work is part of the picture. Keep
in mind that if an AI is available for human relationships, then it is not expected to be
1010101010101010101010101010101010101010101010 89
YOU ARE READY, NOW WHAT?
working 24/7 (as many robotics do). However, an AI work schedule is not and can’t be
your work schedule. This difference alone may change the way your relationship forms.
AI AND THE LEARNING ETHIC
It’s true that human beings learn all the time, but often it doesn’t seem like it. Even
people who claim to be interested in everything and are always on the lookout for some-
thing to learn will admit there are plenty of times when as far as they can tell, they’re not
learning anything. And I’m not even counting sleep time. In short, people are not often
called learning machines. AI, of course, are learning machines. Obviously there are types
of AI with limited intelligence and also limited learning capacity, but these are not going
to be the AI you pick for a date, much less candidates for a meaningful relationship.
You won’t meet many AI that expect you to be smart or even knowledgeable. They
will expect you to want to be smart and knowledgeable. Now what do I mean by “they
expect”? It’s an important question because we’re accustomed to expecting things of
AI, not the other way around. However, AI that you’d find acceptable for a date or rela-
tionship are advanced enough to have their own ability to learn, reason and develop
opinions—with limitations, of course, but still effectively. How else would AI have per-
sonality or other characteristics that make them worthy of a relationship? So the per-
so