AI Creativity and the Human-AI Co-creation Model

Abstract

Artificial intelligence (AI) is bringing new possibilities to numerous fields. There have been a lot of discussions about the development of AI technologies and the challenges caused by AI such as job replacement and ethical issues. However, it’s far from enough to systematically discuss how to use AI creatively and how AI can enhance human creativity. After studying over 1,600 application cases across more than 45 areas, and analyzing related academic publications, we believe that focusing on the collaboration with AI will benefit us far more than dwelling on the competing against AI. “AI Creativity” is the concept we want to introduce here: the ability for human and AI to co-live and co-create by playing to each other’s strengths to achieve more. AI is a complement to human intelligence, and it consolidates wisdom from all achievements of mankind, making collaboration across time and space possible. AI empowers us throughout the entire creative process, and makes creativity more accessible and more inclusive than ever. The corresponding Human-AI Co-Creation Model we proposed explains the creative process in the era of AI, with new possibilities brought by AI in each phase. In addition, this model allows any “meaning-making” action to be enhanced by AI and delivered in a more efficient way. The emphasis on collaboration is not only an echo to the importance of teamwork, but is also a push for co-creation between human and AI. The study of application cases shows that AI Creativity has been making significant impact in various fields, bringing new possibilities to human society and individuals, as well as new opportunities and challenges in technology, society and education.

Full text

AI Creativity and the Human-AI Co-creation
Model
Zhuohao Wu1(B), Danwen Ji2,KaiwenYu
3, Xianxu Zeng4, Dingming Wu5,
and Mohammad Shidujaman6
1School of Animation and Digital Arts, Communication University of China, Beijing, China
2College of Design and Innovation, Tongji University, Shanghai, China
3San Jose State University, San Jose, CA 95112, USA
4Department of Mathematics, University of British Columbia, Vancouver, Canada
5College of Computer Science and Software Engineering,
Shenzhen University, Shenzhen, China
6Department of Information Art and Design, Academy of Arts and Design, Tsinghua
University, Beijing, China
Abstract. Artificial intelligence (AI) is bringing new possibilities to numerous
fields. There have been a lot of discussions about the development of AI technolo-
gies and the challenges caused by AI such as job replacement and ethical issues.
However, it’s far from enough to systematically discuss how to use AI creatively
and how AI can enhance human creativity. After studying over 1,600 application
cases across more than 45 areas, and analyzing related academic publications, we
believe that focusing on the collaboration with AI will benefit us far more than
dwelling on the competing against AI. “AI Creativity” is the concept we want to
introduce here: the ability for human and AI to co-live and co-create by playing to
each other’s strengths to achieve more. AI is a complement to human intelligence,
and it consolidates wisdom from all achievements of mankind, making collabora-
tion across time and space possible. AI empowers us throughout the entire creative
process, and makes creativity more accessible and more inclusive than ever. The
corresponding Human-AI Co-Creation Model we proposed explains the creative
process in the era of AI, with new possibilities brought by AI in each phase. In
addition, this model allows any “meaning-making” action to be enhanced by AI
and delivered in a more efficient way. The emphasis on collaboration is not only
an echo to the importance of teamwork, but is also a push for co-creation between
human and AI. The study of application cases shows that AI Creativity has been
making significant impact in various fields, bringing new possibilities to human
society and individuals, as well as new opportunities and challenges in technology,
society and education.
Keywords: Creativity ·Artificial intelligence ·Design methods and techniques ·
Design process management ·HCI theories and methods ·Education
1 Introduction
In recent years, Artificial intelligence (AI) is bringing new possibilities to numerous
fields from everyday life, industry application to scientific research. There have been a
© Springer Nature Switzerland AG 2021
M. Kurosu (Ed.): HCII 2021, LNCS 12762, pp. 171–190, 2021.
https://doi.org/10.1007/978-3-030-78462-1_13

172 Z. Wu et al.
lot of discussions about the development of AI technologies and the challenges caused
by AI such as job replacement and ethical issues. However, it’s far from enough to sys-
tematically discuss how to use AI creatively and how AI can enhance human creativity.
After studying over 1,600 application cases across more than 45 areas, and analyzing
related academic publications, we believe that focusing on the collaboration with AI will
benefit us far more than dwelling on the competing against AI [1,2]. Among the human
dominant abilities, creativity is one of the most im-portant yet the least understood of
all intellectual abilities until today. It is a popular topic yet remains underdiscussing. It
is being refocused in the era of AI as the debate arose whether AI has creativity.
The purpose of this paper is to introduce the preliminary definition of “AI Creativity”
and the corresponding Human-AI Co-Creation Model. AI Creativity refers to the ability
for human and AI to co-live and co-create by playing to each other’s strengths to achieve
more. AI is a complement to human intelligence, and it consolidates wisdom from
all achievements of mankind, making collaboration across time and space possible. AI
empowers us throughout the entire creative process, and makescreativity more accessible
and more inclusive than ever. The corresponding Human-AI Co-Creation Model we
proposed explains the creative process in the era of AI, with new possibilities brought
by AI in each phase. In addition, this model allows any “meaning-making” action to be
enhanced by AI and delivered in a more efficient way. The emphasis on collaboration
is not only an echo to the importance of teamwork, but is also a push for co-creation
between human and AI.
By illustrating the application cases in various areas, this paper explains that AI
Creativity is a new philosophy to collaborate with all achievements of mankind across
time and space, a new strategy to boost productivity and to inspire innovation, and a new
force to empower human to access creativity inclusively more than ever.
2 Related Work
2.1 The Rise of AI and Potential Impacts
The application of AI has already made significant impacts in businesses around the
world. Between 34% and 44% of global companies surveyed are using AI in their IT
departments mainly in information technology, marketing, finance and accounting, and
customer service, monitoring huge volumes of machine-to-machine activities [3].
AI Industries is becoming the new engine of economic development. According to
recent reports, the potential contribution of AI to the global economy will reach $15.7
trillion, bringing a GDP boost of up to 26% to local economies [4]. 29–62% of annual
growth rate of GVA (gross value added) comes from AI by 2035 across 16 industries in
12 economies, which could lead to an economic boost of US$14 trillion [5].
In these circumstances, people are concerned that AI is on the verge of reaching and
challenging “human-level intelligence,” with recent evidences suggesting that workforce
transitions from human to AI has been triggered. A report by McKinsey [6] says, “50% of
the time spent on work activities in the global economy could theoretically be automated
by adapting currently demonstrated technologies”. The most quoted study of occupations
likely to be automated in the next decade by Oxford University [7] also predicted up to
47% replacement of the workforce in US. A following study by Asian Development Bank

AI Creativity and the Human-AI Co-creation Model 173
[8] concluded that routine, cognitive and manual work would be the most vulnerable,
and the time spent on different activities which can be replaced ranges from 9% to 78%.
2.2 Collaboration Between Humans and AI
AI is relatively strong when it comes to repetitive and predictable workflow, and super
good at dealing with complexity and multi-tasking; while humans are flexible and cre-
ative, and adept at knowledge understanding and strategic thinking, as is summarized in
the figure below. Collaboration between humans and AI varies across domains [9,10].
Human leads where tasks are more about creative or strategy and compassion is needed,
while AI leads where tasks are more about routine or optimization and compassion is
not needed (Fig. 1).
Fig. 1. (a) Human-AI complement each other, (b) Blueprint of Human-AI collaboration [10]
2.3 AI and Creativity
Creativity Research is Mainly About People, Methods and Tools Before the Era of AI.
Creativity is often considered as an “intuition” and can’t be easily interpreted in a ratio-
nal way. The creative industries often refer to graphic design, film, music, video game,
fashion, advertising, media, or entertainment industries [11], related to the extraordinary
thinking by supreme creative individuals [12]. However, creativity actually lies in all
creating activities, from art to science, from everyday life to industry production. And
the thinking behind creativity and all those great creations can be acquired by ordinary
people with deliberate practice [12]. Some recent studies summarized creativity as a
“multifaceted phenomenon to form value and produce innovation entailing the genera-
tion of new intangible or physical item” [13,14]. In discussing how to define, measure,
and enhance the impact of creativity, scholars have proposed three dichotomies: firstly,
whether creativity orginates within individual or comes from social; secondly, whether
creative artifacts should be of novelty or value; and thirdly, whether creative activity is
a thought or an action [15].
In recent decades, great progress has been made by an increasing number of scholars
and researcher with different backgrounds, including cognitive science, psychology,

174 Z. Wu et al.
philosophy, computer science, logic, mathematics, sociology, architecture, design and
etc. The emerging concepts such as digital creativity [16] and computational creativity1
[17] also shed some light on the concept of AI Creativity. With the new possibilities AI
bringing in, the research on AI Creativity will help us use AI creatively and enhance
human creativity efficiently and effectively.
The Creativity of AI is Inseparable from Human’s. From AlphaGo to AlphaFold
[18,19], AI created something which have never existed before,although researchers
still hold different viewpoints on whether AI has creativity [20–23]. Theories and algo-
rithms were invented to imitate and go beyond human’s ways of thinking, using the past
achievements of mankind, such as internet data, as training data for AI. Furthermore, as
lack of understanding causality of the real world, AI has to be designed for interactive
use by humans and enhance human creativity [20,24].NomatterhowfarAIcango
on creativity in the future, human judgment should always be kept essential through the
creative process, so to make sure AI Creativity serves humanity. The creativity of AI
can be considered as a new tool but also beyond a tool.
Human Creativity is to be Enhanced by AI. More and more scholars are studying the
AI-powered, AI-enhanced or AI-assisted human creativity [24,25], reporting the appli-
cation of AI in the creativity industry [26] and Art industries [27]. Designers and design
researchers also discussed and practiced design with data [28,29]. Most of the existing
reports categorize application cases by disciplines or technologies [1,30], while the
report we made, CREO AI Creativity Report 2021, presents the typical cases in the cre-
ative process of our Human-AI Co-Creation Model. Our study shows that AI can work
far more than a black box, but can assists humans throughout the entire creative process.
Human input serves as the framework of this process.
2.4 The Study of Creativity
The Theories of Creativity. One major opinion in early creativity research argued
whether creativity is “unconscious thinking” or “unconscious processing” [31,32], a
sudden appearance of an idea [33] or leaps of insight [34,35]. These theories undoubt-
edly demonstrated the mysterious nature of creativity or the extraordinary thinking of
some creative individuals, which makes ordinary people think that creativity is far away
from them.
Measurement psychology suggested that every individual has creativity, with “diver-
gent thinking” and “convergent thinking” involved [36]. Their study identified the main
components in creative thinking, but lacked the analyzing of the functional mechanisms
of creative process. Guilford [37] further construed creativity as a form of problem-
solving and argues that the creator’s sensitivity to the problem is the key to initiating
creation. In the era of AI, various sensors and big data gives humans expanded views in
both perceptual and rational perspectives.
1Computational creativity also known as artificial creativity, mechanical creativity, creative
computing or creative computation.

AI Creativity and the Human-AI Co-creation Model 175
Evolutionary theories of creativity [38] suggested that the creative process is similar
to Darwinian evolution: Innovative ideas are generated based on early ones and tested
with the latest conditions [35]. The Human-AI Co-Creation Model is consistent with this
theory: Upon input, AI can generate a myriad of explorations and present the preliminary
selections for humans to choose for further “evolving”.
Cognitive theories of creativity believed that the processes of creative thinking and
thinking involved in solving ordinary problems are basically the same [39], and creative
products come through the process of ordinary thinking [12]. This view brings creativity
and ordinary people closer together. With AI empowering in all phases of the creative
process, people can access creativity inclusively more than ever.
In recent years, new theoretical frameworks proposed from the perspective of brain
science have systematically elaborated the interaction between knowledge and creative
thinking [40]. The studies of creative cognition through medical imaging have opened
up new possibilities for the measurement of creativity [41]. It also provides valuable
references for the development of AI.
The Methods of Creativity. These methods encourage creative actions and have
demonstrated their usefulness in both arts and sciences [42], usually covering infor-
mation acquisition, idea generation, problem reframing, prototyping, testing, iterating
and etc. [43].
Some of them are for guiding the creative process, such as TRIZ [44] (and its modi-
fications and derivatives such as SIT and USIT), CPS (Creative Problem Solving) [45],
Design Thinking [46], Double Diamond Model [47], First Principles [48]. Others pro-
vide different thinking principles and toolkits, such as Six Thinking Hats [49], Her-
rmann Brain Dominance Instrument [50], Lateral Thinking, Brainstorming, Brainwrit-
ing, Think Outside the Box, SWOT Analysis, Thought Experience, and Five Ws. As
internet technology advanced, new methods came in such as Data-Driven Design [28],
HEART & GSM [51], Agile Development [52], Design Sprint [53].
Most of the methods listed above are in the context of problem solving. However,
meaning making including painting and music composing, is not necessarily only about
problem solving.
The Abilities of Creativity. Creativity tests summarized the abilities of creativity.
Structure of Intellect theory (SOI) [54] by Guilford organizes intellectual abilities in
three categories: operations, content and products. Operations dimension included six
general intellectual processes. Built on SOI, Torrance Tests of Creative Thinking (TTCT)
[55] involves simple tests of divergent thinking and other problem-solving skills. Sev-
eral controversial creativity tests, such as Getzels and Jackson’s exploration [56] and
Wallach and Kogan’s study [57] had significant impact in the research area (Table 1).
Future-oriented creativity requires people to learn and create in a constantly evolv-
ing technological landscape. ISTE (International Society for Technology in Education)
provided a well-recognized standard for student to become a transformative learner [58].
AIK12 [59] provided a list of criteria of competencies for young people to have in the
era of AI.
From the current study of creativity, some characteristics were found: The classical
creativity theories have less discussion on the discovery strategies before the problem

176 Z. Wu et al.
emerges. The existing creativity methods are mainly for problem solving or product
innovation, but not for other meaning-making activities; Besides, these methods lay
no emphasis on collaboration. The Human-AI Co-Creation Model we proposed in-
troduces new possibilities brought by AI throughout the creative process, allows any
“meaning-making” action to be enhanced by AI and delivered in a more efficient way,
and emphasizes on collaboration no mater it’s interpersonally or between hu-man and AI.
This model also well supports the creativity abilities reflected in the ex-isting creativity
assessment standards such as ISTE and aik12.
Table 1. Main process/components in current study of creativity
Theories/methods/abilities Main process/components
Psychometrics Divergent thinking, Convergent thinking
Evolutionary Theories Randomness, Conditions, Selection
Cognitive Perspective Remembering, Imagining, Planning, Deciding
Brain Research Knowledge Domain: Emotional, Cognitive
Processing Model: Deliberate, Spontaneous
TRIZ (ARIZ) Abstraction, Solution, Concretization
CPS (Creative Problem Solving) Clarify, Ideate, Develop, Implement
Design Thinking Empathize, Define, Ideate, Prototype, Test
Double Diamond Challenge, Discover, Define, Develop, Deliver, Outcome
First Principles Thinking Identify problems and define assumptions, Breakdown the problem into its fundamental
principles, Create new solutions based on the deductions of those principles
Six Thinking Hats White Hat – Facts and Information. Red Hat – Feeling and Intuition. Black Hat – Caution and
Problems. YellowHat – Benefits and Advantages. Blue Hat – Managing Thinking. Green Hat
– Creativity and Solution.
The Whole Brain Thinking Model Analytical Thinking, Structural Thinking, Relational Thinking, Experimental Thinking
Data-Driven Design Goal, Problem/Opportunity Area, Hypothesis, Test, Result
HEART & GSM Happiness, Engagement, Adoption, Retention, Task Success; Goals, Signals and Metrics
Agile Development Requirements, Design, Development, Testing, Deployment, Review
Design Sprint Map, Sketch, Decide, Prototype, Test
SOI’s (Structure of Intelligence)
Operations Dimension
Cognition, Memory recording, Memory retention, Divergent production, Convergent
production, Evaluation
TTCT (Torrance Tests of Creative Thinking) Fluency, Flexibility, Originality, Elaboration
5C Core Competences Cultural Competence, Creativity, Collaboration, Critical Thinking, Communication
ISTE Empowered learner, Digital citizen, Knowledge constructor, Innovative Designer,
Computational Thinker, Creative communicator, Global Collaborator
Human-AI Co-Creation Model Perceive, Think, Express, Collaborate, Build, Test
3 The “AI Creativity”
3.1 Preliminary Definition of AI Creativity
Combining the above points of view, we make our own preliminary definition: AI cre-
ativity is the ability for human and AI to live and create together by playing to each
other’s strengths. It is a new philosophy, a new strategy, and a new force.
AI Creativity can be perceived as a new philosophy. Through AI, people can collab-
orate with all achievements of mankind across time and space. It’s well demonstrated

AI Creativity and the Human-AI Co-creation Model 177
in the making of Portrait of Edmond Belamy: the original algorithm by Americans, the
implementation by Frenchmen, and the AI trained with 15,000 human paintings from
between the 14th and 20th centuries.
AI Creativity can be conceived as a new strategy. Human and AI can play to each
other’s strengths and embrace more possibilities efficiently. Thus human and AI can
complement each other throughout the entire creative process, boosting productivity
and inspiring innovation.
AI Creativity can be regared as a new force. Empowered by AI, human can access
creativity inclusively more than ever. AI creativity can lower the bar to enter an area and
enable human to focus on the most creative part, leaving the complex or time-consuming
tasks to AI.
3.2 AI Creativity Reshapes Creative Process
The “Human-AI Co-Creation Model” is a circular process model including 6 major
phases: perceiving, thinking, expressing, collaborating, building and testing (Fig. 2).
Fig. 2. The Human-AI co-creation model
The first phase is to persive, where human perception can be enhanced by big data
and sensors with AI. Beyond the Senses that humans normally perceive the world with,
AI can turn big data into meaningful information and knowledge using all kinds of
sensors and networks, giving human expanded views in both perceptual and rational
perspectives. The second phase is to think, where humans can think deeper and wider with
AI. Inspiration and exploration that AI brings can go far more than human considerations.
This will break the limits of resource and help human think deeper, wider, in a more
thorough but also efficient way, potentially leading to unexpected accomplishments.
The third phase is to express, where humans can explore more and rapidly with AI.
Various ideas and diverse people need their optimal ways to present, such as painting,
designing, composing, writing, performing, coding, prototyping… Empowered by AI
tools, people won’t be stopped for lack of talent or training. Creativity matters more
than skills. The fourth phase is to collaborate, where human and AI play to each other’s

178 Z. Wu et al.
strengths. Whether working alone or with others, people can always team up with AI.
Just fully understand the strengths and limitations of both human and AI, and give each
side the best assignment. The fifth phase and the sixth are to build and test, where
production can achieve higher quality and lower cost by simulating and analyzing with
AI. Rehearsing gives people a chance to predict how things will go and to prepare
ourselves for real-world events. With detailed simulation and calculation offered by AI,
the process and result of building and testing can be handled effectively and efficiently.
During this creative process, human and AI can complement each other and unleash the
great potential of both sides.
4 The Application of “AI Creativity”
4.1 AI Creativity Prospering Across Industries
We analyzed over 1,600 AI Creativity cases across more than 45 areas from 2017 to
2020. An AI application is only qualified as an AI Creativity case when AI is used
creatively or AI enhances human creativity. Culture and entertainment contributes the
most cases until now, mainly in the format of digital media, which is easier to go with
AI; Cases in industry and lifestyle are trying to bridge the virtual world and the physical
world, which has a great potential to grow; The big percentage in science suggests that
it’s still the early phase of exploration overall; A lot more subcategories will rise from
the misc. as AI moves forward (Fig. 3).
4.2 Examples Throughout the Human-AI Co-creation Model
Perceive. The sound of cellphone-recorded coughs can be used to detect asymptomatic
Covid-19 infections (Fig. 4a). Only AI could achieve high accuracy and efficiency for
this purpose because neither doctor can be effectively trained on this, nor can enough
doctors be trained around the globe for this. It demonstrates the huge potential of AI
assisted diagnosis [60,61].
Sensors such as cameras, LiDAR and millimeter-wave radar on autonomous driving
cars (Fig. 4b), do not only help humans get an all-direction and all-weather view, but
also give humans smart advice based on object detection and analysis [62].
IoT connects machines, objects, animals and people in increasingly numerous ways
(Fig. 4c). As each pig is recognized and traced, customized plans can be applied. Such
a detailed overview enables humans to have a better understanding and greater control
over their work and life [63].
Ambient Intelligence makes physical spaces sensitive and responsive to the presence
of humans (Fig. 4d). It enables more efficient clinical workflows and improved patient
safety in hospital spaces. It could also help the elderly with chronic diseases in daily
living spaces [64].
Think. Predicting the protein structure, AlphaFold unlocked a greater understanding
of what it does and how it works (Fig. 5a). From AlphaGo, AlphaStar to AlphaFold,
AI demonstrated new methods and great potential to learn and solve complex problems
effectively and efficiently through massive exploration [19].

AI Creativity and the Human-AI Co-creation Model 179
Fig. 3. AI creativity cases across industries: (a) By name of subcategories, (b) By number
Multi-Channel human-machine interaction allows humans to communicate with AI
in a natural way [68], such as searching by color and shape (Fig. 5b). Making AI adapt
to the human way, it brings not only comfort, but also efficiency [65].
Inspired by the knowledge graph based on the search queries on internet, people can
get a better overview of the object studied and can trigger more relevant ideas around
it (Fig. 5c). With the help of AI, the world’s knowledge can be organized and accessed
more than ever, and then further developed into new concepts [66].
Simulating a simple game of hide-and-seek, agents built a series of distinct strategies
and counterstrategies, some of which were unexpected (Fig. 5d). This further suggests
extremely complex and intelligent behaviors could be synthesized [67].

180 Z. Wu et al.
Fig. 4. (a) Cough test for Covid-19 by MIT [60]. (b) Lidar and camera view at night by Waymo
[62]. (c) Pig recognition and smart farming by JD.com [63]. (d) Illuminating the dark spaces of
healthcare with ambient intelligence by Li Fei-Fei et al. [64]
Express. It’s not a dream anymore to have a “Magic brush” that turns a doodle into a
photo. As if done by an experienced artist or designer, AI turns people’s rough ideas into
reality (Fig. 6a). AI helps humans focus more on generating and testing ideas without
worrying about the presenting skills [69]. And DALL·E released in Jan 2021 is pushing
this to the next level.
AI can work as a friend sharing ideas with humans, responding and inspiring each
other to make a story gradually (Fig. 6b). From such a game today, we can foresee that
the future of collaborative writing between human and AI is coming [70].
With the help of AI, humans can play any role in any context by controlling char-
acters through body and face movements (Fig. 6c). The making of animations and
demonstrations becomes easier [71].
Coding has a history of becoming easier to use, and AI will speed up this process
(Fig. 6d). Although the making of high-quality software still requires experienced engi-
neers, it will be world-changing to enable everyone to create their own software by
talking, writing, drawing, playing with building blocks, not only by coding [72].

AI Creativity and the Human-AI Co-creation Model 181
Fig. 5. (a) AlphaFold on protein folding problem by DeepMind [19]. (b) Street art by color
by Google Arts & Culture [65]. (c) Knowledge graph of google searches by Anvaka [66]. (d)
Simulation of multi-agent hide and seek by OpenAI [67]
Build and Test. In the context of product manufacturing, AI allows designers and engi-
neers to input their design goals, along with parameters such as materials, manufacturing
methods, and cost constraints (Fig. 7a). Then AI explores all the possible solutions by
testing and iterating [73].
Qualitative changes can happen when quantitative changes are big enough. Person-
alization comes after. Alibaba Luban’s design engine has demonstrated how powerful
the true personalization is, as so does TikTok’s recommendation engine (Fig. 7b). AI is
the key to enable massive design and implementation efficiently at low cost [74].
A process that takes generations of evolution in the physical world can be simulated
in the virtual world at much higher speeds. Through the design and making of Xenobots,
biology and computer scientists worked together and significantly speeded up the process
of trial and error (Fig. 7c) [75].
Digital Twin brings parallel universe to reality. Building and testing happens in the
virtual world and the best solution can be chosen to implement in the physical world
(Fig. 7d). Meanwhile, anything manifesting in the physical world can be reflected in the
virtual world for further analysis and exploration [76].

182 Z. Wu et al.
Fig. 6. (a) GauGAN by NVIDIA [69]. (b) AI Dungeon by Latitude [70]. (c) Animation production
with Kuaishou & PuppetMaster [71]. (d) Build apps by describing in words with debuild, powered
by GPT-3 [72]
Collaborate. We placed collaborating at the end instead of as its sequence in the model,
because it’s a great example demonstrating the creative process enhanced by AI. Art
styles are among the great achievements of civilizations. It’s almost impossible for a
human to master every style of art, but it’s not hard for AI (Fig. 8a). Trained with
examples of various artistic styles, AI can imitate any one of them. Based on the variations
developed upon the input, the best parts can be picked for further development2.
Expressing in an ancient language is much harder than understanding it. However,
given enough training materials, it’s no different for AI to learn a modern language or
an ancient one (Fig. 8b). Such a poetry AI can give human many inspirations, although
it’s not perfect, nor does it have a soul3.
2The painting by Mr. HOW with Deep Dream Generator: https://deepdreamgenerator.com/.
3The poem Mr. HOW with Tsinghua JiuGe: http://118.190.162.99:8080/, Microsoft JueJu: http://
couplet.msra.cn/jueju/ and SouYun: https://sou-yun.cn/MAnalyzePoem.aspx.

AI Creativity and the Human-AI Co-creation Model 183
Fig. 7. (a) “The first chair created with AI” by Philippe Starck, Autodesk & Kartell [73]. (b)
Alibaba Luban’s AI banner design [74]. (c) Xenobots: first living robots by University of Vermont
[75]. (d) Collaborative robots in assembly by University of Southern Denmark [76]
Breaking language barriers, AI helps regular people to enjoy different cultures and
create things more easily (Fig. 8c) [77]. In this case, the poem in ancient Chinese was
firstly translated into modern Chinese by human, then into English by AI, and then
fine-tuned by human in the end4.
Not everyone can make music although it’s a universal language (Fig. 8d). Music AI
allows people to bring out the rhythm and rhyme from their heart and mind. Based on
the initial input, variations will be generated for picking for further developing5.
As the ending of a traditional Chinese painting, a stamp was used, which is actually
a QR code linked to the video of the making of this artwork6.
4The poem translated by Mr. HOW with Google, Apple and Microsoft Translation.
5The music by Mr. HOW with LingDongYin: https://demo.lazycomposer.com/compose/v2/.
6The making of the artwork, The Mind of AI Creativity: http://qr09.cn/Ew06EW.

184 Z. Wu et al.
Fig. 8. The mind of AI creativity (a) Painting, (b) Poem in Chinese, (c) Poem in English, (d)
Music composing
5 The Future of AI Creativity
5.1 Developing AI Creativity
The adoption of AI is expected to run high across industries, company sizes and geog-
raphy [78]. Skills of using AI as a demand in all online job vacancies have been
rapidly increasing especially since 2015 [79]. AI talents include technology developers,
technology-product transformers [80], and product utilizers. All of them need AI Cre-
ativity, mastering AI thinking and skills. Regardless of interests or specialties, people
can always find effective ways to develop their AI Creativity.
Collaborative creation between human and AI will be seen everywhere. Processes are
enhanced or even evolved in every step where AI enters. STEM-DAL [81] is a new way
to inspire and leverage AI Creativity in cross-disciplinary learning and creation. Science
and Art stand at the two ends. Technology brings Science into application, while Design
brings Art into application. Meanwhile, Engineering merges Technology and Design,
while Mathematics and Literature serve as foundations (Fig. 9).
In recent years, some educational initiatives have proposed some new concepts and
practices for AI education in addition to coding, such as AI4ALL [82], aik12-MIT [59]
and Mr. HOW AI Creativity Academy [81]. These initiatives are trying to make AI
education inclusively accessible for all people with different interests and specialties,
beyond technology perspectives only. Great potential remains yet to be unleashed.

AI Creativity and the Human-AI Co-creation Model 185
Fig. 9. The STEM-DAL model
5.2 Challenges and Opportunities
Technology Perspective. Pre-Mature AI technologies are still the norm, although there
has been a huge leap since the modern era of deep learning began at the 2012 ImageNet
competition. Scientists are working hard on the next generation of AI to break today’s
constrains. New initiatives such as GPT-3 aren’t perfect, but it did unleash new potentials.
Incomplete solutions built on separate technologies according to proprietary stan-
dards are very common. Doing anything with AI often relies heavily on experts and well-
funded organizations. This stands in the way of more people adopting AI technologies
and unleash the great potential of AI Creativity.
Limited Resources invested in the industry today are mainly driven by capital for to
maximize financial returns. As more scientists, engineers and other resources are added
to transform AI technologies into more products, more areas will be covered and driven
by AI Creativity.
Attacks on AI raises the alert of AI security. A simple attack can cause AI to see
something different than what an image really looks like, so a T-shirt with a special
pattern can render someone “invisible” to cameras, or an autonomous car misread an
altered traffic sign. Addressing the security issue also needs AI Creativity.
Society Perspective. Privacy could become very vulnerable in the era of AI. Individ-
uals could become transparent to all kinds of ubiquitous sensors and AI applications
with recognition and analysis abilities. It’s not always easy to balance privacy with
convenience. AI Creativity may help to make smart decisions.
Abuse of AI has been documented, such as unwanted facial recognition, spam calls,
deepfake videos, etc. Anyone of those could set you in trouble. As more people access
more AI applications, potential for abuse will increase. AI Creativity can contribute in
preventing this.
Discrimination follows from human behaviors, as AI is trained with materials that
humans generated. As a technical issue, things like accuracy of facial recognition across
different human races can be easily improved. However, things like the bias in resume
screening may need a lot of AI Creativity to address.
Job Replacement is a hot topic, especially for parents. However, many parents
wouldn’t necessarily want their children to do the jobs AI is going to replace. Rather
than worrying about potential job replacement, it’s more important to think about how

186 Z. Wu et al.
to inspire and develop AI Creativity for both yourself and future generations, and how
to live and create with AI.
Education Perspective. Exploration in the AI education system is key for developing
talents. What to learn, how to train, whether AI should be an independent discipline or
intertwined with other ones… The answers to these questions will only emerge through
deep thinking and practice. AI Creativity itself shall be used in these explorations.
Liberal Arts +AI education is almost an untouched area, when compared with
STEM, robotics and coding. As part of AI Creativity, the fusion of liberal arts and AI
will be critical. It’s not only to give all kids a balanced education in addition to STEM
learning, but also to provide a method of development for kids who do not have an
aptitude for STEM.
Thinking vs Skills are both important things to learn for AI Creativity. AI thinking
is for choosing the right strategy, while skills are for choosing the right tactics. They
support each other and can’t live without each other.
Democratizing AI and Creativity will be a key outcome of AI Creativity education.
People will be empowered to go beyond their current level, to live and create with AI
by playing to each other’s strengths.
6 Conclusion
AI Creativity has been making significant impact in various fields, bringing new pos-
sibilities and challenges to human society and individuals. The topic of cultivating AI
Creativity has a great value and potential to be explored. For professionals, the evolved
thinkings and methods need to be built up, such as the Human-AI Co-Creation Model
we proposed; For educators, the inclusive AI education system needs to be built up, such
as the STEM-DAL system we proposed, which will be discussed in another paper; For
scholars, the framework of measuring AI Creativity needs to be built up. For everyone,
how to unleash the great potential of AI Creativity and how to prevent the unexpected
consequences need to be discussed. We initiated CREO (Creativity Renaissance in Edu-
cation and Organizations) with experts around the world, and will keep pushing the
boundary of AI Creativity.
Acknowledgements. The authors like to thank Qinwen Chen, Guojie Qi, Peiqi Su, Qing Sheng,
Jieqiong Li, Qianqiu Qiu, Linda Li and all the volunteers for their contribution in this paper.
References
1. Mueller, S.T., Hoffman, R.R., Clancey, W., Emrey, A., Klein, G.: Explanation in Human-AI
Systems: a Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography
for Explainable AI (2019)
2. Peeters, M.M.M., van Diggelen, J., van den Bosch, K., et al.: Hybrid collective intelligence in
a human-AI society. AI Soc. 36, 217–238 (2021). https://link.springer.com/article/10.1007/
s00146-020-01005-y

AI Creativity and the Human-AI Co-creation Model 187
3. Tata Consultancy Services: How Companies are Improving Performance using Artificial
Intelligence (2017)
4. PwC: Sizing the prize. What’s the real value of AI for your business and What’s the real value
of AI for your business and how can you capitalise? (2017)
5. Accenture: How AI Boosts Industry Profits and Innovation (2017)
6. McKinsey Global Institute: Jobs Lost, Jobs Gained. Workforce Transitions in A Time of
Automation (2017)
7. Frey, C.B., Osborne, M.A.: The Future of Employment. How Susceptible Are Jobs to
Computerisation (2013)
8. Asian Development Bank: Asian Development Outlook. Asian Development Bank, Manila,
Philippines (2018)
9. Amershi, S., et al.: Guidelines for human-AI interaction. In: CHI 2019 Proceedings of the
2019 CHI Conference on Human Factors in Computing Systems, Glasgow, Scotland UK, 4–9
May, 2019. ACM, New York (2019)
10. Lee, K.F.: How AI can save our humanity. https://www.ted.com/talks/kai_fu_lee_how_ai_
can_save_our_humanity
11. Howkins, J.: The Creative Economy: How People Make Money from Ideas. Penguin, UK
(2002)
12. Weisberg, R.W.: Creativity: Understanding Innovation in Problem Solving, Science, Inven-
tion, and the Arts. John Wiley (2006)
13. Palmiero, M., Piccardi, L., Nori, R., Palermo, L., Salvi, C., Guariglia, C.: Editorial: creativity:
education and rehabilitation. Front. Psychol. 10, 1500 (2019)
14. Wikipedia: Creativity. https://en.wikipedia.org/w/index.php?title=Creativity&oldid=999
304631
15. Kaufman, J.C., Sternberg, R.J. (eds.): The Cambridge Handbook of Creativity. Cambridge
University Press, Cambridge (2019)
16. Owen Kelly: Digital creativity (1996)
17. Colton, S., Wiggins, G.A.: Others: computational creativity: the final frontier? ECAI 12,
21–26 (2012)
18. OpenAI: AlphaFold : a solution to a 50-year-old grand challenge in biology. https://deepmind.
com/blog (2020)
19. Deepmind: AlphaFold: a solution to a 50-year-old grand challenge in biology. https://dee
pmind.com/blog/article/alphafold-a-solution-to-a-50-year-old-grand-challenge-in-biology
20. Boden, M.A.: Creativity and artificial intelligence. Artif. Intell. 103, 347–356 (1998)
21. Erden, Y.J.: Could a created being ever be creative? Some philosophical remarks on creativity
and AI development. Minds Mach. 20, 349–362 (2010)
22. Fetzer, J.H., Dartnall, T.: Artificial Intelligence and Creativity. An Interdisciplinary Approach.
Springer Netherlands, Dordrecht (2010).
23. Dartnall, T. (ed.): Artificial Intelligence and Creativity. An Interdisciplinary Approach.
Springer Netherlands, Dordrecht (1994)
24. Miller, A.I.: The Artist in the Machine. The World of AI-Powered Creativity. The MIT Press,
Cambridge, Massachusetts (2019)
25. Fischer, G., Nakakoji, K.: Amplifying designers’ creativity with domain-oriented design envi-
ronments. In: Dartnall, T. (ed.) Artificial Intelligence and Creativity. An Interdisciplinary
Approach, pp. 343–364. Springer Netherlands, Dordrecht (1994)
26. Anantrasirichai, N., Bull, D.: Artificial Intelligence in the Creative Industries: A Review
(2020)
27. Chen, W., Shidujaman, M., Tang, X.: AiArt: Towards Artificial Intelligence Art (2020)
28. King, R., Churchill, E.F., Tan, C.: Designing with Data: Improving the User Experience with
A/B Testing. O’Reilly Media, Incorporated (2017)

188 Z. Wu et al.
29. Graham Michael Dove: CoDesign with data (2015)
30. JSAI AI Map task force: AI Map Beta (English), 6 June 2019
31. Poincaré, H.: The Foundations of Science (GH Halstead, Trans.). Science , New York (1913)
32. Miller, A.I.: Insights of Genius: Imagery and Creativity in Science and Art. Springer Science
& Business Media (2012)
33. Wertheimer, M.: Max Wertheimer Productive Thinking. Springer (2020). https://doi.org/10.
1007/978-3-030-36063-4
34. Csikszentmihalyi, M.: Flow and the Psychology of Discovery and Invention, vol. 39.
HarperPerennial, New York (1997)
35. Simonton, D.K.: Origins of Genius: Darwinian Perspectives on Creativity. Oxford University
Press (1999)
36. Mumford, M.D.: Something old, something new: revisiting Guilford’s conception of creative
problem solving. Creat. Res. J. 13, 267–276 (2001)
37. Guilford, J.P.: Varieties of creative giftedness, their measurement and development. Gifted
Child Q. 19, 107–121 (1975)
38. Campbell, D.T.: Blind variation and selective retentions in creative thought as in other
knowledge processes. Psychol. Rev. 67, 380 (1960)
39. Newell, A., Shaw, J.C., Simon, H.A.: The processes of creative thinking. In: Contemporary
Approaches to Creative Thinking, 1958, University of Colorado, CO, US; This paper was
presented at the aforementioned symposium (1962)
40. Dietrich, A.: The cognitive neuroscience of creativity. Psychon. Bull. Rev. 11, 1011–1026
(2004)
41. Arden, R., Chavez, R.S., Grazioplene, R., Jung, R.E.: Neuroimaging creativity: a psychome-
tric view. Behav. Brain Res. 214, 143–156 (2010)
42. Chen, W., Shidujaman, M., Jin, J., Ahmed, S.U.: A methodological approach to create inter-
active art in artificial intelligence. In: Stephanidis, C., et al. (eds.) HCII 2020. LNCS, vol.
12425, pp. 13–31. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-60128-7_2
43. Wikipedia: Creativity techniques. https://en.wikipedia.org/w/index.php?title=Creativity_t
echniques&oldid=996672803
44. Al’tshuller, G.S.: The Innovation Algorithm: TRIZ, Systematic Innovation and Technical
Creativity. Technical innovation center, Inc. (1999)
45. Creative Education Foundation: The CPS Process. https://www.creativeeducationfoundation.
org/creative-problem-solving/the-cps-process/
46. Brown, T.: Others: design thinking. Harv. Bus. Rev. 86, 84 (2008)
47. Design Council: What is the framework for innovation? Design Council’s evolved Dou-
ble Diamond. https://www.designcouncil.org.uk/news-opinion/what-framework-innovation-
design-councils-evolved-double-diamond
48. Merrill, M.D.: First principles of instruction. Educ. Tech. Res. Dev. 50, 43–59 (2002)
49. de Bono, E.: Six Thinking Hats. Penguin,s UK (2017)
50. Herrmann: HBDI (Herrmann Brain Dominance Instrument)® Assessment | Herrmann. https://
www.thinkherrmann.com/hbdi
51. Rodden, K., Hutchinson, H., Fu, X.: Measuring the user experience on a large scale: user-
centered metrics for web applications. In: Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems, pp. 2395–2398 (2010)
52. Abrahamsson, P., Salo, O., Ronkainen, J., Warsta, J.: Agile Software Development Methods:
Review and Analysis (2017)
53. Banfield, R., Lombardo, C.T., Wax, T.: Design Sprint. A Practical Guidebook for Building
Great Digital Products. O’Reilly Media, Ind, Sebastopol (2016)
54. Guilford, J.P.: The structure of intellect. Psychol. Bull. 53, 267 (1956)
55. Torrance, E.R.: The torrance tests of creative thinking. Norms-technical mannual (1990)

AI Creativity and the Human-AI Co-creation Model 189
56. Getzels, J.W., Jackson, P.W.: Creativity and intelligence: explorations with gifted students.
AAUP Bull. 48, 186 (1962)
57. Wallach, M.A., Kogan, N.: Modes of Thinking in Young Children: A Study of the Creativity-
Intelligence Distinction. Holt Rinehart and Winston (1965)
58. ISTE: ISTE Standards for Students. https://www.iste.org/standards/for-students
59. k12 AI Education MIT: aik12-MIT. https://aieducation.mit.edu/
60. MIT: MIT Covid-19 Initiative. https://opensigma.mit.edu/
61. Laguarta, J., Hueto, F., Subirana, B.: COVID-19 artificial intelligence diagnosis using only
cough recordings. IEEE Open J. Eng. Med. Biol. 1, 275–281 (2020). https://doi.org/10.1109/
OJEMB.2020.3026928
62. Waymo: Waymo Open Dataset: Sharing our self-driving data for research. https://blog.
waymo.com/2019/08/waymo-open-dataset-sharing-our-self.html
63. JD.com: Pig Facial Recognition – YouTube. https://www.youtube.com/watch?v=BECDKK
Yi-48
64. Johnson, J., Alahi, A., Fei-Fei, L.: Perceptual losses for real-time style transfer and super-
resolution. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol.
9906, pp. 694–711. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46475-6_43
65. Google Arts & Culture: Color Explorer: Street Art - Google Arts & Culture. https://artsandcu
lture.google.com/color?project=street-art&col=RGB_1BE5E1
66. Kashcha, A.: Autocomplete VS graph. https://anvaka.github.io/vs/?query=creativity
67. Baker, B.: Emergent Tool Use from Multi-Agent Interaction. OpenAI (2019)
68. Shidujaman, M., Zhang, S., Elder, R., Mi, H.: “RoboQuin”: a mannequin robot with natural
humanoid movements. In: 2018 27th IEEE International Symposium on Robot and Human
Interactive Communication (RO-MAN), pp. 1051–1056 (2018)
69. Park, T., Liu, M.-Y., Wang, T.-C., Zhu, J.-Y.: Semantic image synthesis with spatially-adaptive
normalization. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern
Recognition, pp. 2337–2346 (2019)
70. Latitude: AI Dungeon. https://play.aidungeon.io/
71. Mr. HOW: Animation with Facial Recognition & Motion Capture on Mobile Phone. https://
v.kuaishou.com/7reyAJ
72. Debuild: Build web apps lightning fast. https://debuild.co/
73. STARCK Site web officiel: A.I. - Introducing The First Chair Created With Arti-
ficial Intelligence. https://www.starck.com/a-i-introducing-the-first-chair-created-with-artifi
cial-intelligence-p3801
74. Alibaba Cloud Community: The Evolution of Luban in Designing One Bil-
lion Images. https://www.alibabacloud.com/blog/the-evolution-of-luban-in-designing-one-
billion-images_596118
75. University of Vermont: Team Builds the First Living Robots. https://www.uvm.edu/uvmnews/
news/team-builds-first-living-robots
76. University of Southern Denmark: Collaborative Robots in Assembly: A Digital-Twin
Approach. https://www.youtube.com/watch?v=lWHT3_B2spg
77. Shidujaman, M., Mi, H.: “which country are you from?” A cross-cultural study on greeting
interaction design for social robots. In: International Conference on Cross-Cultural Design,
pp. 362–374 (2018)
78. Ransbotham, S., Kiron, D., Gerbert, P., Reeves, M.: Reshaping Business with Artificial
Intelligence. Closing the Gap Between Ambition and Action (2017)
79. Alekseeva, L., Azar, J., Gine, M., Samila, S., Taska, B.: The demand for AI skills in the labour
market. https://voxeu.org/article/demand-ai-skills-labour-market

190 Z. Wu et al.
80. Wang, K.-J., Shidujaman, M., Zheng, C.Y., Thakur, P.: HRIpreneur thinking: strategies
towards faster innovation and commercialization of academic HRI research. In: 2019 IEEE
International Conference on Advanced Robotics and its Social Impacts (ARSO), pp. 219–226
(2019)
81. Mr. HOW AI Creativity Academy: Inspire Kids to Live and Create with AI by Playing to
Each Other’s Strength. http://qr09.cn/CVUI2H
82. AI4ALL: AI4ALL Annual Report 2019 (2019)