ArticlePDF Available

Pervasive persuasive technology and environmental sustainability (Workshop)

Authors:

Christine Satchell

Queensland University of Technology

Eric Paulos

University of California, Berkeley

Show all 5 authorsHide

Accepted proposal for a workshop held at the 6th International Conference on Pervasive Computing, May 19th, 2008, Sydney, Australia

Image MoreAssociates

…

Screenshot from the I Am Green Facebook application

…

Overall system behaviour – a cycle of sensing/monitoring resource usage, reasoning about the way the resource is used and acting

…

Figures - uploaded by Christine Satchell

Content may be subject to copyright.

Content uploaded by Christine Satchell

Content may be subject to copyright.

QUT Digital Repository:

http://eprints.qut.edu.au/

Foth, Marcus and Satchell, Christine and Paulos, Eric and Igoe, Tom and Ratti,

Carlo (2008) Pervasive persuasive technology and environmental sustainability

(Workshop). In: 6th International Conference on Pervasive Computing, 19 May

2008, Sydney.

Conference(Workshop(Proceedings(

Pervasive(Persuasive(Technology(and(Environmental(

Sustainability(

!"#$%&"'(&)*+(,-(."-/0-.1,"-(2,1&(3)#45%,4)(6778(

9:1&(;5<(6778=(><+-)<=(?>!=(@0%1#5*,5(

Workshop(website(

&11'ABB222C4#"*,$C+)BD#))-B(

Workshop(organisers,(and(editors(of(the(proceedings(

• ;5#.0%(E"1&(FG0))-%*5-+(H-,4)#%,1<("I(J).&-"*"D<=(@0%1#5*,5K(

• L&#,%1,-)(>51.&)**(FG0))-%*5-+(H-,4)#%,1<("I(J).&-"*"D<=(@0%1#5*,5K(

• M#,.(350*"%(FN-1)*(O)%)5#.&(P)#$)*)<=(H>@K(

• J"Q(ND")(F?RH=(H>@K(

• L5#*"(O511,(F;NJ=(H>@K(

Accepted(papers(

• ?"1)%("-(1&)(3"*,1,.5*(NQ5D)A(3)#45%,4)(L"Q'01,-D=(;"+)*,-D(@%%)QS*5D)=(

5-+(M."*"D,.5*(T"4)#-5-.)C(P)-/5Q,-(UC(P#511"-(9=(?515*,)(V)#)Q,/)-$"(6(

9(H-,4)#%,1<("I(L5*,I"#-,5=(W"%(@-D)*)%=(H>@(

6(?)2(R"#$(H-,4)#%,1<=(H>@(

• J&)(3"1)-1,5*("I(HS,L"Q'(J).&-"*"D,)%(1"(X)1)#Q,-)(1&)(L5#S"-(

E""1'#,-1%("I(3#"+0.1%C(@*,(X5+5(9=Y=(J&"#%1)-(>155$)(6=(M*D5#(E*),%.&(9=6(

9(N-%1,101)("I(J).&-"*"D<(;5-5D)Q)-1(FNJM;ZU>TK=(H-,4)#%,1<("I(>1C(

T5**)-=(>2,1[)#*5-+(

6(N-I"#Q51,"-(;5-5D)Q)-1=(MJU(\]#,.&=(>2,1[)#*5-+(

Y(>@3(O)%)5#.&(LML(>1C(T5**)-=(>2,1[)#*5-+(

• J"25#+%(35#1,.,'51"#<(X)%,D-("I(@QS,)-1(3)#%05%,4)(J).&-"*"D<C(V5-)1(

X54,%C(X)'5#1Q)-1("I(L"Q'01)#(>.,)-.)=(T#,--)**(L"**)D)=(H>@(

• ;"1,451,-D(M-4,#"-Q)-15**<(>0%15,-5S*)(P)&54,"#(L&5-D)%(2,1&(5(^,#105*(

3"*5#(P)5#C(J525--5(X,**5&0-1=(T)"I(P).$)#=(V)--,I)#(;5-$"II=(O"S)#1(

_#501C(U0Q5-ZL"Q'01)#(N-1)#5.1,"-(N-%1,101)=(L5#-)D,)(;)**"-(

H-,4)#%,1<=(3,11%S0#D&=(H>@(

• 3",-1%("I(3)#%05%,"-A(>1#51)D,.(M%%)-1,5*,%Q(5-+(M-4,#"-Q)-15*(

>0%15,-5S,*,1<C(350*(X"0#,%&C(X"-5*+(P#)-(>.&""*("I(N-I"#Q51,"-(5-+(

L"Q'01)#(>.,)-.)%=(H-,4)#%,1<("I(L5*,I"#-,5=(N#4,-)=(H>@(

• H-+)#%15-+,-D(Q"1,451,"-(5-+()-5S*,-D(5.1,"-(1"25#+%(.&5-D)C(3)--<(

U5D)-=(X0-.5-(H-+)#2""+C(+,D,15*)%$,Q"C-)1=(>0##<(U,**%=(?>!=(@0%1#5*,5(

• J&)(?)2(!)**ZJ)Q')#)+(M-4,#"-Q)-1A(J0-,-D(P0,*+,-D%(5-+(L,1,)%C(X5-(

U,**=(X0-.5-(!,*%"-C(5#0'C."Q=(W"-+"-=(H_=(5-+(><+-)<=(@0%1#5*,5(

• 3#"Q"1,-D(M-4,#"-Q)-15**<(>0%15,-5S*)(P)&54,"#%(H%,-D(>".,5*(;5#$)1,-D(

,-(MQ)#D,-D(3)#%05%,4)(J).&-"*"D,)%C(`Q5#(_&5-=(V"&-(L5--<C(P)#$)*)<(

N-%1,101)("I(X)%,D-=(L"Q'01)#(>.,)-.)(X,4,%,"-=(H-,4)#%,1<("I(L5*,I"#-,5=(

P)#$)*)<=(H>@(

• L"-1)a1Z@25#)(3)#45%,4)(3)#%05%,4)(><%1)Q%(I"#(;5-5D,-D(!51)#(5-+(

M-)#D<(H%5D)=(5-+(L`6(MQ,%%,"-%A(;0*1,ZW)4)**)+(3"*,.,)%=(T"5*%=(5-+(5-(

Ma')#1(><%1)Q%(>&)**(@''#"5.&C(>)-D(!C(W"$)=(V0D+011(>,-D&=(U5,(W)C(W5(

J#"S)(H-,4)#%,1<=(;)*S"0#-)=(@0%1#5*,5(

• H%,-D(')#%05%,4)(1).&-"*"D<(1"()-."0#5D)(%0%15,-5S*)(S)&54,"#C(L))%(

;,++)-=(J)++<(;.L5**)<=(V55'(U5Q=(O00+(\55*S)#DC(M,-+&"4)-(H-,4)#%,1<(

"I(J).&-"*"D<=(J&)(?)1&)#*5-+%(

• 35#1,.,'51)A(3#"+0.,-D(5(;5%%(>.5*)(M-4,#"-Q)-15*(L5Q'5,D-(I"#(

3)#45%,4)(J).&-"*"D<C(;5#$(35a1"-C(;,a)+(O)5*,1<(W5S=(H-,4)#%,1<("I(

?"11,-D&5Q=(H_(

• J&)(X)%,D-("I(NQ'#,-1A(b!5*$(1&)(!5*$b(5-+(`1&)#(W)%%"-%C(\5.&5#<(

3"0%Q5-=(U5I)[(O"0[51,=(_51,)(L"**,-%=(V"&-(>15%$"C(T#5'&,.%=(

^,%05*,[51,"-(c(H%5S,*,1<(L)-1)#=(T)"#D,5(J).&=(@1*5-15=(H>@((

• H-+)#%15-+,-D(5-+(N-I*0)-.,-D(>'51,"ZJ)Q'"#5*(^,%,1"#(;"4)Q)-1(,-(

?51,"-5*(35#$%(P5%)+("-(>151,.(5-+(X<-5Q,.(>)-%"#(X515C(_51/5(

>.&).&1-)#=(U)*Q01(>.&#"QZE),)#15DC(5#%)-5*C5.C51=(^,)--5=(@0%1#,5(

• J)##53)+A(@(X)%,D-(L"-.)'1(I"#(>0%15,-5S,*,1<C(V)--,I)#(>1),-C(>.&""*("I(

L,-)Q51,.(@#1%=(H-,4)#%,1<("I(>"01&)#-(L5*,I"#-,5=(W"%(@-D)*)%=(H>@(

• L&5**)-D,-D(L"QI"#1(c(L*)5-*,-)%%(?"#Q%(1&#"0D&(N-1)#5.1,4)(N-ZU"Q)(

E))+S5.$(><%1)Q%C(R"*5-+)(>1#)-D)#%C(>.&""*("I(T*"S5*(>10+,)%=(>".,5*(

>.,)-.)(5-+(3*5--,-D=(O;NJ(H-,4)#%,1<=(;)*S"0#-)=(@0%1#5*,5(

• J5$,-D(1&)(T0)%%2"#$("01("I(M-4,#"-Q)-15**<(>0%15,-5S*)(W,I)%1<*)%C(

O"-5$(>015#,5(9=(@5*"$(X)%&Q0$&(6(

9(0#S5-%)-%"#%C."Q=(>5-(E#5-.,%."=(H>@(

6(O".$<(;"0-15,-(N-%1,101)=(P"0*+)#=(L`=(H>@(

• M."N%*5-+A(@(><%1)Q(I"#(3)#%05+,-D(H%)#%(1"(O)+0.)(L`6(MQ,%%,"-%C(

L&,&,#"(J5$5<5Q5(9=(^,*,(W)&+"-4,#15(6(

9(!5%)+5(H-,4)#%,1<=(J"$<"=(V5'5-6(U)*%,-$,(N-%1,101)(I"#(N-I"#Q51,"-(

J).&-"*"D<=(E,-*5-+(

Workshop on

Pervasive Persuasive

Technology and

Environmental Sustainability

ORGANIZERS:

Marcus Foth (Queensland University of Technology, Australia)

Christine Satchell (Queensland University of Technology, Australia)

Eric Paulos (Intel Research Berkeley, USA)

Tom Igoe (NYU, USA)

Carlo Ratti (MIT, USA)

PERVASIVE PERSUASIVE TECHNOLOGY AND

ENVIRONMENTAL SUSTAINABILITY

Marcus Foth, Christine Satchell, Eric Paulos,

Tom Igoe, Carlo Ratti

m.foth@qut.edu.au, satc@unimelb.edu.au, eric@paulos.net,

tom.igoe@nyu.edu, ratti@mit.edu

www.urbaninformatics.net/green

Abstract

Environmental sustainability and climate change are issues which must no longer be ignored by

anyone, any industry or any academic community. The pervasive technology, ubiquitous computing

and HCI community is slowly waking up to these global concerns. The key theme of this workshop

around environmental sustainability will be addressed threefold: (1) How to go beyond just

informing and into motivating and encouraging action and change. (2) Pervasiveness can easily

turn invasive. We want to start re-considering the impact of pervasive technology from an

ecological perspective. (3) Digital divide between humans and the environment: Can the process of

‘blogging sensor data’ assist us in becoming more aware of the needs of nature? How can we

avoid the downsides?

1. Theme of the workshop and topics of interest

Environmental sustainability and climate change are issues which must no longer be ignored by

anyone, any industry or any academic community. The pervasive technology, ubiquitous

computing and HCI community is slowly waking up to these global concerns.

The Nobel Peace Price 2007 was awarded to Al Gore and the Intergovernmental Panel on Climate

Change (IPCC) “for their efforts to build up and disseminate greater knowledge about man-made

climate change, and to lay the foundations for the measures that are needed to counteract such

change”. The citation highlights the urgency of the fact that information and awareness around

causes and implications are necessary but not sufficient to combat climate change. Action is

required.

The key theme of this workshop around environmental sustainability will be addressed threefold:

1. Providing people with environmental data and educational information – via mass

communications such as film, TV and print and new media, or micro communications such as

pervasive sensor networks (cf. Participatory Urbanism and Ergo at urban-atmospheres.net; real-

time Rome at senseable.mit.edu; biomapping.net; placeengine.com) – may not trigger sufficient

motivation to get people to change their habits towards a more environmentally sustainable

lifestyle. This workshop seeks to develop a better understanding how to go beyond just

informing and into motivating and encouraging action and change.

2. Pervasiveness can easily turn invasive. It has already caused negative consequences in

biological settings (e.g., algae in lakes and oceans, kudzu vine in the southeastern US, rabbits

and cane toads in Australia). Pervasive can be a dangerous term when the ecological impacts

are disregarded. Pervasive technology is no different. In order to avoid further serious damage

to the environment, this workshop aims to lay the foundations to start re-considering the impact

of pervasive technology from an ecological perspective.

3. Addressing the 21st century Digital Divide: The mass uptake of pervasive technology brings

about digitally networked and augmented societies; however, access is still not universal.

Castells and others use the notion of the ‘digital divide’ to account for those whose voices are

not heard by this technology. Initially, the divide was seen only between the first and third

worlds and then between urban and rural, but with today’s near ubiquitous coverage, the digital

divide between humans and the environment needs to be addressed. Virtual environments

could give the natural world an opportunity to ‘speak’. How can we address imbalances? For

example, sensors embedded in the environment could allow creeks and rivers to blog their own

pollution levels, local parks can upload images of native bird life. Can the process of ‘blogging

sensor data’ (sensorbase.org) assist us in becoming more aware of the needs of nature? How

can we avoid the downsides?

2. Topics of interest

Topics of interest include but are not limited to:

• Transfer persuasive and motivational approaches and experiences from design cases which

successfully employ pervasive technology in areas such as games and entertainment, health,

and marketing and advertising, e.g., competition, collaboration, rewards, team play, make it fun.

• Innovative ways or re-appropriated ways to reduce the impact of computing production (e.g.,

increase the life cycles of computing devices; re-purpose older computing devices for sensor

networks, data logging and other low-performance but increasingly useful tasks; re-think the

design of computing devices to allow for more efficient and thorough recycling of

components).

• Considerations of what ‘pervasive technology’ means from an ecological perspective.

• New applications of pervasive computing technology to support environmental education and

decision making in formal (school, work) and informal (leisure, play, everyday) settings.

• Evaluations and evaluation methods for assessing the impact of pervasive computing devices,

applications on the environment.

• New interfaces of pervasive computing devices, systems and applications and modes of

interactions between people and nature.

3. Format

After the introduction of the organisers and the key themes of the workshop, we will do a little

icebreaker activity in the form of mini interviews followed by brief informal peer introductions of

all participants. We then want to generate a common knowledge base for the workshop on

environmental sustainability by collecting information on what we know about the issues at hand,

e.g., sources of data on climate change, accessibility and legibility of that data, current impact it has

or lack thereof. After morning tea, the workshop breaks into three rotating groups (starting with

1/2/3) according to the three themes: 1. Motivation, 2. Ecological impact, and 3. Digital divide

between humans and the environment. After lunch, these groups rotate themes so each group works

on each theme: 2/3/1, followed by 3/1/2. Following afternoon tea, we hold a plenary to share results

and discuss further steps, plan the Design Challenge 2009, discuss collaboration arrangements

beyond the workshop and allocate tasks to volunteers.

Design Challenge 2009: Impact! In addition to the conventional academic outcomes, we want to

define a number of feasible goals and design a process to bring these goals to fruition within 12

months time and present them at Pervasive 2009. Additionally, we want the workshop to be an

opportunity to exchange research insights, expertise and ideas. We also want to leave enough

breakout and social time to allow for professional networking opportunities.

4. The organisers

Marcus Foth: ARC Australian Postdoctoral Fellow, Institute for Creative Industries and

Innovation, Queensland University of Technology, Brisbane, Australia and 2007 Visiting Fellow,

Oxford Internet Institute, UK. Interests: urban informatics, master-planned communities, social

computing, social networks, triple bottom line sustainability, wombats. urbaninformatics.net

Christine Satchell: ARC Australian Postdoctoral Fellow (Industry), Institute for Creative

Industries and Innovation, Queensland University of Technology, Brisbane, Australia. Interests:

HCI, scenario and persona design, young people, digital nomads, mobile technology, puppies.

Eric Paulos: Director, Urban Atmospheres, Intel Research Berkeley, USA. Interests: urban

computing, social telepresence, robotics, physical computing, interaction design, persuasive

technologies, intimate media, sasquach. paulos.net

Tom Igoe: Associate Arts Professor, Interactive Telecommunications Program, Tisch School of the

Arts, New York, USA. Interests: physical computing and networking, sensor networks,

ecologically sustainable practices in technology development, monkeys. tigoe.net

Carlo Ratti: Director, SENSEable City Laboratory, MIT, Boston, USA. Interests: Architecture,

sensor networks, real-time data, urban technology, urban planning, crocodiles. senseable.mit.edu

5. A selection of relevant readings

CLIMATE RISK. (2007). Towards a High-Bandwidth, Low-Carbon Future: Telecommunications-

based Opportunities to Reduce Greenhouse Gas Emissions. Fairlight: Climate Risk Australia.

DARBY, S. (2006). The Effectiveness of Feedback on Energy Consumption (DEFRA report).

Oxford, UK: Environmental Change Institute, University of Oxford.

DE YOUNG, R. (2000). Expanding and Evaluating Motives for Environmentally Responsible

Behavior. Journal of Social Issues, 56(3), 509-526.

FOGG, B. J. (2003). Persuasive Technology: Using Computers to Change What We Think and Do.

Amsterdam: Morgan Kaufmann Publishers.

FOGG, B. J., & ECKLES, D. (2007). Mobile Persuasion: 20 Perspectives on the Future of

Behavior Change. Stanford, CA: Stanford Captology Media.

HERRING, H. (1998). Does Energy Efficiency Save Energy: The Implications of accepting the

Khazzoom-Brookes Postulate. Retrieved Apr 2, 2008, from

http://technology.open.ac.uk/eeru/staff/horace/kbpotl.htm

MONBIOT, G. (2006). Heat: How to Stop the Planet From Burning. Cambridge, MA: South End

Press.

PAULOS, E., HONICKY, R. J., & HOOKER, B. (2008, in press). Citizen Science: Enabling

Participatory Urbanism. In M. FOTH (Ed.), Urban Informatics. Hershey, PA: IGI Global.

Notes on the Political Image:

Pervasive Computing, Modeling Assemblage,

and Ecological Governance

Abstract:

What are the larger opportunities for pervasive computing technologies to monitor

and model the intricate assemblages of the natural and artificial environments?

Perhaps more importantly, how might such media contribute to or even constitute in

a new kind of

reflexive governance

in the image of the information they produce?

We consider what is at stake for such a constitution by locating it in theoretical

insights from other contexts. In outlining current and future research directions, we

underscore the necessity of environmentally produced information to concretize

itself not merely as data for some later, deferred political action, but as a direct

political image: an

instrumental diagram

in its own right.

Agencies: Borders, Sensors, Interfaces Far from borderless, our world is filled with

an apparent infinity of political borders , biological borders, logistical borders,

informational borders. “Things” (flora, fauna, and machines, data) assemble,

expunge and express themselves within this segmented landscape.

We see such borders also as

interfaces

, membranes which govern the

conditions of exchange between any paired complex bodies (from ambient air and

soil to a bank customer and her money.)

An environmental sensor, be it a flower petal or inscribed silicon wafers

microcasting in near-field communication, is, in Bruno Latour’s parlance, a kind of

actant

. (Latour, 2005) It speaks on behalf of some condition in this little network and

communicates to other parts of the system. The interface, in this case that sensor,

thus takes the position of a limited agent in the whole system, and in it, the smallest

transactional unit of data, becomes itself the emergent actor in this landscape of

borders.

The Pervasive Scenario: To Govern What? The driving scenario understood by us,

is one in which there would be a local administration of each interface’s expression --

what toxins flowers absorb, what plastics are allowed through customs, what concrete

infrastructure cracks under slow pressure—based on what is heard and is relayed

through a pervasive network of networks of computational monitoring media.

But once data is gathered what does this image look like in detail, and what is

done with it once we have it?

Individual events in this listening landscape are absorbed as signifiers, are

correlated according to as yet unimagined protocols, and are registered not only as

information about which a governance might act, but for us as the actual medium of

some non-governmental political apparatuses would register themselves back onto

the world.

That is, first the information is a map or diagram and then, more interestingly,

that same information becomes an interface.

In this, governance is not delivered by the management of “human rights,”

but instead acts according to new constitutional forms, yet to be formally ratified. Its

work is not simply policing an environmental homeostasis, but of direct management

by participation of ecosystemic interfaces

as the constitution of a new polis and

politics.

(Mouffe, 2005)

It is that imminent political diagram, not yet present, that redefines interfaces

less as pure information than as positions in an expanded parliamentary territory;

neither a ‘standing reserve’ of objective datapoints nor a undifferentiated naturalized

unity. (Heidegger, 1993)

Policy and Relationality At NYU’s Environmental Health Clinic, these operations

are designed as a matter of institutional policy (or policy is made as a matter of

design.) They are a model for how ecological interfaces, both human and non-

human, both organic and inorganic, can be understood as site of

health

. And health

is then less the individual body and the medicines that might be inserted into it to

contain undesirable states, than it is the external, plural, at-hand living and non-living

worlds of the urban landscape. Here a fertile middle-ground is taken as the location

of prescription and operation. The individual lung and the gathering storm system

are engaged at once by design, as both public health records in their ways.

Both provide an open form of evidence of what the administrative unit and

ecological microcosm that is NYU is doing and has done through its multi-scalar

participations in the urban ecology of the city. As with any Clinic, at EHC

prescriptions are given. But here prescriptions are offered both to individuals and to

architectural systems on how to better monitor and adjust or redesign the causal

interrelationships in which they work.

Some Recent Projects Toward This: A Portfolio of Prescriptions Some projects at

NYU’s EHC work to monitor and measure that evidence, and others to redesign

through it.

For example, One Tree Project, in which genetically identical trees were

planted across the Bay Area acts as a formidable monitoring interface for the effects

of ambient environmental variation on the complex mechanisms of organic growth.

In another,, 1400 face masks were distributed to the mingling hordes of protesters

and pedestrians during the Republican convention in New York, which together

worked as a networked surface on which individual consumption of air pollutants

was traced and tracked. In these, distributed interfaces express and even compute,

socially legible evidence of ecological interactions. But instead of such symptoms

rendered through a mediating layer of silicon computation, these simple, extremely

analog devices draw their evidence more directly.

This allows a more direct experiential response to the information they

express, but like any such image they require another step before becoming

instruments of change and governance.

Other NYU EHC projects directly redesign the

assemblages

at work

(DeLanda, 2006). No Park, for example, uses a legal gap in traffic storage interfaces

(i.e. no parking zones in front of hydrants) to install intricately designed gardens

which absorb automobile related pollutants right where they are generated. Here the

socio-ecological network of the city is neither smothered nor rarified, but amplified

and engaged by configuring it at the level (and height) of collective assembly.

Future projects seek to engage at the scale of the NYU micro-city and its

architectures. One locates solar panels not on rooftops but as window shades,

circumventing the difficult regulations of solar collection technologies in Manhattan

as well as locating the production side of shared energy infrastructure at the direct

personal level of individual dorm and office dwellers. He and she can see and relate

to their energy consumption footprint at the literal scale of his or her own footprint.

Another improves on the vogue for rooftop gardens as decorative micro-parks for

people and turns them into gray water treatment machines that support the

interlocking purposes of ground-level and migratory species across the city.

Data Smog and the Missing Expert In these projects a problem of translation and

activation is introduced, one that is not solved by the amassing of more

computational power and the scattering of sensors into the world. That is, natural

systems work very well already as monitoring devices, but we have yet to formulate

complete constitutional images of their participation. We have yet to find the best

ways of making an

image-instrument

of their political profiles, diagrams of their

evidence that constitute an effective agency.

Scaled somewhere between the medical care of the individual body and the

continental sweep weather modeling , the an ecological

polis

is largely unmanaged

and unarticulated.

To introduce an layer of engineered listening and speaking media into and

onto the ambient environment (in parking lots, wetlands, cloud clusters, lung cells,

etc.) will open a flood of information about how worldly systems perform and relate.

By flipping the on-switch, as it were, and being at once able to monitor and model

the interrelationships of all assemblages at once, would be akin to the invention of

the microscope, opening up the complexity and agency of worlds we could not

imagine.

But would it also be a din of voices that we simply do not have the means to

properly listen to, to govern through: a churning cacophony of signals?

The design impulse will be to model these signals into dataclouds, undulating

traffic clusters that allow datasets to be sliced and figured by different patterns and

variables. We believe that such data clouds have purpose and potential but cannot

by themselves realize either. Dataclouds have a tendency to inappropriately reassure

their audiences. They imply, in the grace and intricacy of their renderings, a

presumed expert system (and expert) on whose behalf they are fictitiously designed

who

must

be using these information as an instrument of government, somewhere. If

not now, soon. Dataclouds signify control but too often only defer agency to that

missing expert. This is part of the work that they do as blog culture memes. They

assure; they present the

affect

of a political agency that is still to find its bearings.

The data cloud looks like expert instrumentation, but does the closed frame

of its bounded diagnostic, in fact make it easier for those viewing the data cloud to

retract their own involvement back into logic of simulation and to remain spectators?

What is the political space of air quality? What would it mean for such images to

work more effectively as interfaces? How can the images of evidence, produced by

pervasive sensing media, work not just to display information about socio-

ecosystmes but to turn their audiences into users who can, in the direct course of

their habitation of the world, compute by their interactions a preferred assemblage

of what the image represents? How can the data that becomes diagrams, become

again instruments of a new political space?

The answer to these, we believe, lies in the specification of both an

ecologically appropriate political scale and a constitutional image of that

polis

, which

in turn may

rely on the encounters between pervasive computing and ecological

governance.

Notes:

Manuel De Landa,

A New Philosophy of Society: Assemblage Theory and Social

Complexity

. Continuum. London. 2006.

Martin Heidegger, “The Question Concerning Technology”

Basic Writings

. Ed.

David Krell. HarperCollins. New York. 1993 (1954)

Bruno Latour,

Reassembling the Social: An Introduction to Actor-Network Theory.

Oxford University Press. 2005

Chantal Mouffe,

On the Political

. Routledge, New York. 2005.

Author Biographies

Benjamin H. Bratton

(SCI_Arc, UCLA, Yahoo!) invents systems concepts and

translates and transposes them into actual systems. This labor requires him to where

many hats for different occasions, including sociologist, design strategist, professor,

software executive, and historian of exceptional violence. He teaches architects about

double-bind ironies at SCI_Arc, media artists about topologies of logistics at UCLA,

and enterprise product strategists about the social specification of emergent data

channels at Yahoo!.

Natalie Jeremijenko

(NYU Environmental Health Clinic) is a polymath design

technologist and political affectician, working within the wormholes connecting

experimental art and global science policy. She has taught at the world’s august

institutions, been shown in the toniest arts festivals, received the most competitive

awards, and annoyed the most miscast authoritarians.

THE POTENTIAL OF UBICOMP TECHNOLOGIES TO

DETERMINE THE CARBON FOOTPRINTS OF

PRODUCTS

Ali Dada1,3, Thorsten Staake2, Elgar Fleisch1,2

1 Institute of Technology Management (ITEM-HSG), University of St. Gallen, Dufourstrasse 40a,

9000 St. Gallen, Switzerland

2 Information Management, ETH Zurich, Sonneggstrasse 63, 8092 Zurich, Switzerland

3 SAP Research CEC St. Gallen, Blumenbergplatz 9, 9000 St. Gallen, Switzerland

ali.dada@unisg.ch, tstaake@ethz.ch, elgar.fleisch@unisg.ch

www.bits-to-energy.ch

Abstract

Several problems exist in accurately quantifying the greenhouse gases (GHG) which result from the

production, transportation, usage, and recycling of products. We review in this paper the current

attempts at measuring GHG emissions and investigate the potential of UbiComp technologies in

improving the state of the art in carbon footprint calculation and in communicating the result to

consumers. This has potential benefits on consumer awareness and behaviour and can also foster

competition among companies towards higher energy and resource efficiency.

1. Introduction

Companies measure energy usage and greenhouse gas (GHG) emissions to comply with regulations,

to assess their performance in an energy-constrained economy, and to participate in a growing

carbon market [8] Moreover, some organizations also identified the value of communicating the

carbon footprint1 of their products to end consumers who increasingly consider environmental

aspects in their purchasing decisions [4]. However, a number of challenges need to be addressed

when determining GHG emissions for individual products or services. In particular, it is difficult to

capture all relevant emissions along the product’s lifecycle and to account for the variations in

footprints of individual instances of products, such as spatial, temporal, and supplier variations. In

this contribution, we propose a solution that leverages UbiComp technologies in order to

dynamically track a product’s carbon footprint along its life-cycle and make it easily accessible to the

consumer. The benefit of such a solution is that it enables comparisons on many levels, for example

between variations of the same product, between products from different manufacturers, and

between different companies upstream in the value chain. Consumers can then exert pressure by

selecting the least carbon-intensive products. We provide a review of methods used to measure

GHG emissions in section 2 and outline in section 3 where UbiComp technologies can help. We

conclude in section 4.

1 “The carbon footprint is a measure of the total amount of CO2 emissions that is directly and indirectly accumulated

over the life stages of a product.”[11]

2. State of the art in measuring energy usage and GHG emissions

We review in this section the approaches used to determine the emissions of companies (supply side)

and tools that calculate emissions due to the consumption by individuals (demand side).

2.1. Supply-side approaches

There are several standards and guidelines used by companies in order to quantify and report their

GHG emissions. The most prominent standard is the GHG corporate protocol [8], used by the

majority of the FT500 companies and serving as the basis for most GHG guidelines and programs.

This protocol is accompanied by tools that help companies to calculate their emissions. Various

other tools exist in the context of GHG reporting programs such as CARROT for participants in the

California Climate Action Registry [3] and SEIT which is provided with the DoE 1605b voluntary

reporting program [6]. Furthermore, software vendors offer enterprise solutions for carbon and

energy management, e.g. Environmental Compliance by SAP [9].

There has recently been some effort to calculate the carbon footprint of some particular products

using a supply-chain-wide approach. Prominent examples are the pilots conducted by Carbon Trust

together with brands such as Walkers and Trinity Mirror [4]. Furthermore, CarbonCounted provides

an online application for brand owners, also for the purpose of calculating a product’s footprint [5].

These two examples differ in many aspects which makes it impossible to accurately calculate and

compare a product’s footprint, such as whether they include all GHG emissions or only CO2 and

whether they consider offsets. They do not take the energy usage at the retailer into consideration,

or do so in a very coarse-grained way. Neither takes home usage into account. These issues make it

impossible to accurately calculate and compare a product’s footprint. Finally, when different

suppliers have different footprints, or when there are temporal or spatial variances between different

instances of a product, average numbers are used, which results in inaccurate results. Examples

include fruits bought off-season which require six months of chilled storage or products being

shipped half way across the globe.

2.2. Demand-side Approaches

Many carbon calculators are available to determine the carbon footprint of an individual or a

household. According to Bottrill [1], the calculators give an annual result based on one data entry

per activity, thus not taking seasonal or lifestyle fluctuations into consideration. The author adds that

the tools are falling short of accurately monitoring people’s energy use and providing the feedback

required. In the field of monitoring energy usage, there are several projects that provide accurate

numbers by conducting measurements on household devices. For example, Kuckuck is a project

which uses sensor data to display domestic energy consumption [10]. The Device-Level Power

Monitoring system [7] comprises monitoring units that plug into power outlets enabling per-device

electricity monitoring. Despite such projects that measure the energy usage of different devices at

home, this usage is not meant to be attributed to the products that require them, for example the

energy used by a washing machine or a refrigerator is not distributed, respectively, among the

clothes and foods inside. In effect, the energy usage at home cannot be accumulated to the total

carbon footprint of a product which until now is focusing only on the supply-side calculations. In the

next section we will describe how products can have one dynamic carbon footprint which is easily

accessible to users.

3. The potential of UbiComp

To calculate a dynamic carbon footprint, we need a mechanism that links information about

emissions with the products that required them, namely by assigning to each item its share of the

emissions of each emitting process. This can be achieved via the unique identification of items, e.g.

via the Electronic Product Code (EPC) [2]. Items with an EPC number can be tracked from

manufacturing along the supply chain and until the retailers, with the possibility to add dynamic

information – such as the carbon emitted at each partner – to the EPC Information Services (EPC

IS). With unique identification in place, the share of yet-disregarded life-cycle stages of the

product’s footprint can be quantified and attributed to the items that required them. For example, the

same EPC number can be used to identify items beyond the point of sale, where intelligent devices

can attribute their consumed energy to the items using it. Unique identification can also make the

product footprint easily accessible, e.g. to a consumer on his mobile phone. After including all the

product life-cycle stages, we would be able to consolidate the emissions and assign them to

individual items or other entities. This approach therefore comprises two stages. The first stage is an

information gathering stage that can be realized by a publish/subscribe system in which process

owners publish the process information and users (including items) can be subscribed to the

processes they are part of. The second stage is an on-demand calculation of an item’s footprint based

on the available process/item information. Our initial research indicates that only minor changes are

necessary to use EPC Information Systems for the purpose of carbon footprint management.

4. Conclusion

We reviewed in this paper the available approaches in measuring GHG emissions and noted their

shortcomings in contributing to the accurate quantification of carbon footprints of products. We then

highlighted the potential of UbiComp technologies in contributing to this effort. Such technologies

can also make the carbon footprint of products conveniently accessible to consumers. This enables

consumers to exert pressure on the brand owner to decrease the footprint and it helps influence their

own behavior at home in order to keep the footprint as low as possible. Also, if companies can

compare the carbon-intensity of products from different suppliers, they can exert pressure by

selecting partners with the least carbon-intensive products.

References

[1] BOTTRILL, C. Internet-based Tools for Behaviour Change. In European Council for Energy Efficient Economics

(ECEEE) Summer Study 2007: Dynamics of Consumption. 2007.

[2] BROCK, D. The Electronic Product Code (EPC): A Naming Scheme for Objects. Tech. Rep. MIT-AUTOID-WH-

002, MIT Auto ID Center, 2001.

[3] CALIFORNIA CLIMATE ACTION REGISTRY. Carrot. Online http://www.climateregistry.org/, 2008.

[4] CARBON TRUST. Carbon Footprints in the Supply Chain: The Next Step for Businesses, 2006.

[5] CARBONCOUNTED. Online http://www.carboncounted.com/, 2008.

[6] ENERGY INFORMATION ADMINISTRATION. Energy-related Emissions Data & Environmental Analyses. Online

http://www.eia.doe.gov/environment.html, 2007.

[7] LEBLANC, J. Device-Level Power Consumption Monitoring. In Workshop on Ubiquitous Sustainability, UbiComp

(2007).

[8] RANGANATHAN,J.,CORBIER,L.,BHATIA,P.,SCHMITZ,S.,GAGE,P., AND OREN, K. The Greenhouse Gas Protocol -

A Corporate Accounting and Reporting Standard. WBCSD and WRI, 2004.

[9] SAP. SAP Environmental Compliance. Online http://www.sap.com/solutions/grc/ehscompliance/index.epx, 2008.

[10] STRINGER,FITZPATRICK,CHALMERS,HARRIS,KRISHNA, AND HAARLANDER. Kuckuck – Exploring Ways of

Sensing and Displaying Energy Consumption Information in the Home. In Workshop on Ubiquitous

Sustainability, UbiComp (2007).

[11] WIEDMANN, T., AND MINX, J. A Definition of 'Carbon Footprint'. ISA Research Report 07-01, 2007.

Motivation

This paper comprised a brief review of measures to calculate carbon emissions and possible ways to

make carbon footprints more accurate by including the various product lifecycle stages. Since our

focus is on providing the consumers with the appropriate information to empower them in the

decision-making process, the key question is which information consumers need, how to retrieve it,

and how consumers will be influenced by it. Also important is the optimal medium and form of

presenting the information and how to measure their feedback including possible third-order effects.

We are eager to discuss such questions in the workshop and learn from others’ experiences. More

information about us and other related research can be found at www.bits-to-energy.ch.

Authors’ Biographies

Ali Dada

Ph.D. candidate at the Institute of Technology Management in the University of St. Gallen and

research associate at SAP Research. Ali’s Ph.D. topic is carbon footprints of products: their

feasibility and implications on consumer behavior and brand owner’s sustainability decisions.

Dr. Thorsten Staake

Thorsten Staake works as Project Manager of the Bits to Energy Lab and Senior Researcher at the

Chair of Information Management, Swiss Federal Institute of Technology (ETH) Zurich. Before

joining the ETH, Thorsten spent one year at MIT’s Auto-ID Lab and worked for two years at the

Institute of Technology Management, University of St. Gallen, where he was in charge for the

Institute’s supply chain security initiative. His current research interests are related to enhancing

supply chain visibility to determine, communicate, and minimize greenhouse gas emissions.

Prof. Dr. Elgar Fleisch

Professor of Information and Technology Management; Department of Management, Technology,

and Economics, Swiss Federal Institute of Technology (ETH Zurich) and Institute of Technology

Management, University of St. Gallen (HSG); Co-Chair of Auto-ID Labs, Insurance-Lab (I-Lab),

and Bits to Energy Lab. Elgar Fleisch’s current research focuses on the economic and information

system perspective of ubiquitous computing, work that addresses the architecture of the Internet of

Things. He also conducts research in the areas of service engineering and internationalization.

Towards Participatory Design of Ambient Persuasive

Technology

Janet Davis

Dept. of Computer Science

Grinnell College

Grinnell, IA 50112

davisjan@cs.grinnell.edu

1. Design Problem and Method

At my institution, the EcoCampus committee is charged with developing approaches to promote sus-

tainable behavior on campus, speciﬁcally behaviors that will reduce the campus’s net carbon emis-

sions. Approaches that provide new opportunities, such as coordinating a campus rummage sale to

reduce end-of-year trash disposal and start-of-year purchases, have met with great success. More di-

rect approaches to persuade individuals to change their behavior—for example, posting pleas against

food waste and reminders to turn off lights—have had less clear results, and even engendered some

resentment.

I aim to engage this committee and others in participatory design of new persuasive technology aimed

at changing these types of behaviors in ways that are more demonstrably effective, and engaging

rather than annoying. Participatory Design (or PD) is a family of theories and methods related to end

users as full participants in the design process. Many, but not all, PD researchers and practicioners

are motivated in part by a belief in the value of democracy in the design setting [12]. Participatory

design is well-suited to this design context because my institution has a culture of active stakeholder

involvement in decision-making, and because there is a group of people already committed to envi-

ronmental sustainability. I believe it is only right to include them in the design process. By making

them full participants, I gain access to their special knowledge, perspectives, and creativity, as well as

an enthusiastic group of supporters of the new approaches.

The design process will begin with a Future Workshop [9, 10] to guide EcoCampus committee mem-

bers and other interested members of the community in generating visions for behavioral and cultural

change on campus. A Future Workshop consists of three phases: a Critique phase to elicit problems

with current practice, a Fantasy phase to envision an ideal world in which those problems were solved,

and an Implementation phase in which the aim is to begin making plans for realistic changes. Whereas

most Future Workshops in the PD literature have focused on work practices, the workshop concept

originated in the context of community organizing around societal issues [9]; the topic of this work-

shop will be “Green Culture at Grinnell.” At the end of this workshop, I hope we will have identiﬁed

a number of behaviors to target for change, as well as possible approaches—technological and non-

technological—for changing them. I plan to continue working with a smaller group of stakeholders

on one or more behaviors that seem appropriate to the development of persuasive technology.

2. Ambient Persuasion

Designing technology to change behavior with respect to environmental sustainability, or to promote

sustainable culture, clearly falls into B.J. Fogg’s deﬁnition of persuasive technology: “interactive

computing systems designed to change people’s attitudes and behaviors” [4, p. 1]. Indeed, Fogg

uses the example of paper recycling by an organization to illustrate his different perspectives on the

roles of persuasive technology [3]. Several efforts have begun to explore persuasive technology in

relation to environmental sustainability. For example, RideNow is a web- and email-based system

that facilitates ad-hoc ride sharing [15]. The SmartTrip tool for mobile devices simpliﬁes the task of

combining multiple errands into a single trip, with the goal of reducing driving among those resistant

to ride sharing or public transit [8]. GreenScanner is a mobile application that shoppers can use at

the store to read reviews of the environmental impacts of various products [14]. In these systems, the

computer’s functional role is that of a persuasive tool [3, 4]: The system makes suggestions, provides

information, or makes the desired behavior easier to do. But, users must go to some effort to adopt

these tools. As Fogg argues, persuasive technology for mobile devices are most effective when they

help people to achieve the goals they have already decided upon [4, pp. 192–3].

An alternative is to embed persuasive tools in the built environment, particularly in public or semi-

public places. The idea is to make suggestions at exactly the right time and place, without annoying

those to be persuaded [4, 11, 7]. Recent work by Mathew and his students exploits this approach,

which he calls environmental persuasion [11], in the context of promoting physical activity. In one

design, an attractive glass staircase with embedded information displays entices commuters in a sub-

way station to use the stairs rather than the escalator [11]. In another, information kiosks at bus stops

suggest that walking will result in earlier arrival time at the destination than waiting for the bus, and

will burn more calories [13]. Embedding persuasive technology in public spaces allows for incremen-

tal peruasion: “Persuasion is initiated by the persuasive elements, but the actual behavior change is a

result of gradual but increasing awareness of the importance of that change” [11]. Such an approach

can inﬂuence the behavior of those who are not initially committed to behavior change. This gradual

approach also leaves room for discovery, playfulness, ambiguity, and subtlety, qualities that could

make a persuasive device intriguing rather than annoying.

Some design efforts have already begun to explore environmental persuasion with respect to envi-

ronmental sustainability. For example, WaterBot aims to reduce water consumption by tracking and

displaying information about water use at the sink itself [1]. Although the design is aimed at the

home or workplace, some ideas could be adapted to sinks in more public spaces. Intriguingly, Hol-

stius et al. use live and robotic plants in an ambient display that shows the balance between trash

and recycling in a dining area [5]. The goals of ambient displays—to provide awareness through the

physical environment, without demanding attention [16]—seem particularly compatible with the idea

of non-annoying, incremental persuasion. These initial efforts show this area, which one might call

ambient persuasive technology, is ripe for further investigation.

3. Research Questions

I believe that applying participatory design to persuasive technology is a novel approach. Futher-

more, as Hornecker et al. argue, the application of participatory design methods to the design space

of pervasive computing is still largely unexplored [6]. A key question, then, is how (or whether) to in-

corporate known principles and guidelines for the design of persuasive technology and for persuasion

in general, such as Cialdini’s six fundamental human tendencies [2], with the process of participatory

design. A second question is how to incorporate concern for the ethics of persuasion into the process.

Design participants are not the only ones who will interact with the persuasive technology, and they

are not necessarily typical stakeholders as they almost certainly value environmental sustainability

more highly than most community members. Fogg outlines a method to account for ethical concerns

in the design of persuasive technology [4], with which participants might be engaged. Furthermore,

although participatory design is not inherently problematic from the standpoint of environmental sus-

tainability, can particular design themselves be made more sensitive to that goal?

WaterBot [1] and the “infotropism” display [5] not only make indirect suggestions, but also monitor

and report on behavior in the surrounding environment. Because of the community context, such

displays tread a ﬁne line between self-monitoring, which seems generally positive and acceptable,

and surveillance, which Fogg warns may cause public compliance without private acceptance [4, p.

49]. What design features are needed to avoid that undesirable outcome? What guidance can I draw

from these examples? For example, where WaterBot aims to support social validation in a household

context by tracking the water consumption of individuals, the infotropism display, designed for a

more public context, does not connect the data it gathers with particular individuals.

Fogg argues that behavioral change is a more compelling metric than attitudinal change for measuring

the success of persuasive technology: it is thought to be more difﬁcult to achieve, it can be measured

without relying on self-reports, and ﬁnally, it is a direct measure of real-world outcomes [3]. However,

following the work of Holstius, et al. [5] and other work on ambient displays, and consistent with the

goal of not annoying people, it will be important to assess users’ attitudes towards and understanding

of the persuasive technology itself. In the context of environmental sustainability, one should also ask

about the net environmental impact of the persuasive technology. Unlike applications for mobile de-

vices already in use, such as cell phones, ambient persuasive technology involves introducing new de-

vices into the environment and thus involves some measurable consumption of resources—electricity

to power the device and other resources for its manufacture. Can we demonstrate that the devices we

build have an environmental impact that is lesser than that of the undesirable behavior they are in-

tended to change? What techniques can we use to reduce the power consumption of ambient displays,

both in the prototype phase and in deployment? At one extreme, we might eschew the “technology”

aspect of persuasive technology altogether, and use Intille’s approach of temporarily deploying sensor

systems to measure the behavioral impacts of low-technology persuasive techniques [7].

4. Motivation for Attending the Workshop

I have been interested in environmental sustainability and ambient displays for several years; the

theme of “Persuasive Pervasive Technology and Environmental Sustainability” lets me bring those

interests together. At the workshop, I hope to meet other researchers interested in related areas and

perhaps raise some different perspectives on design methodology. I hope also for feedback and new

insights to guide my work, which is in its earliest phases.

5. Biography

Janet Davis is Assistant Professor of Computer Science at Grinnell College, a socially-conscious

liberal arts college in central Iowa. She earned her B.S. in Computer Science at Harvey Mudd College,

and her Ph.D. in Computer Science and Engineering at the University of Washington. Her dissertation

work involved the design of new user interfaces for a large-scale urban simulation system, with a

particular attention to the values of democracy and freedom from bias. She is a member of the

Sustainable CHI group. Her interests include Value Sensitive Design, Participatory Design, ambient

displays, environmental sustainability, design for local impact, and alpacas.

References

[1] E. Arroyo, L. Bonanni, and T. Selker. Waterbot: Exploring feedback and persuasive techniques at the

sink. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI 2005),

pages 631–639, 2005.

[2] R. B. Cialdini. The science of persuasion. Scientiﬁc American, pages 76–81, February 2001.

[3] B. Fogg. Persuasive computers: Perspectives and research directions. In Proceedings of the SIGCHI

Conference on Human Factors in Computing Systems (CHI 1998), pages 225–232, 1998.

[4] B. Fogg. Persuasive Technology: Using Computers to Change What We Think and Do. Morgan Kauf-

mann, San Francisco, 2003.

[5] D. Holstius, J. Kembel, A. Hurst, P.-H. Wan, and J. Forlizzi. Infotropism: Living and robotic plants as

interactive displays. In Proceedings of the 5th Conference on Designing Interactive Systems (DIS 2004),

pages 215–221, 2004.

[6] E. Hornecker, J. Halloran, G. Fitzpatrick, M. Weal, D. Millard, D. Michaelides, D. Cruickshank, and

D. D. Roure. UbiComp in opportunity spaces: Challenges for participatory design. In Proceedings of the

Participatory Design Conference (PDC ’06), 2006.

[7] S. S. Intille. Ubiquitous computing technology for just-in-time motivation of behavior change. In Ubi-

Health Workshop, 2003.

[8] E. M. Johnson, H. Koh, J. McAtee, and S. Shoulders. SmartTrip: Persuasive technology to promote

conscious driving habits. In CHI ’07 Extended Abstracts on Human Factors in Computing Systems, 2007.

[9] R. Jungk and N. M ¨

uller. Future Workshops: How to Create Desirable Futures. Institute for Social

Inventions, London, 1987.

[10] F. Kensing and K. H. Madsen. Generating visions: Future workshops and metaphorical design. In

J. Greenbaum and M. Kyng, editors, Design at Work: Cooperative Design of Computer Systems, chapter 8,

pages 155–168. Lawrence Erlbaum Associates, Hillsdale, NJ, 1991.

[11] A. P. Mathew. Using the environment as an interactive interface to motivate positive bheavior change

in a subway station. In CHI ’05 Extended Abstracts on Human Factors in Computing Systems, pages

1637–1640, 2005.

[12] M. J. Muller. Participatory design: The third space in HCI. In J. A. Jacko and A. Sears, editors, The

Human-Computer Interaction Handbook, chapter 54, pages 1051–1068. Lawrence Erlbaum Associates,

Mahwah, NJ, 2003.

[13] V. Singh and A. P. Mathew. WalkMSU: An intervention to motivate physical activity in university stu-

dents. In CHI ’07 Extended Abstracts on Human Factors in Computing Systems, pages 2657–2662, 2007.

[14] B. Tomlinson. Prototyping a community-generated, mobile device-enabled database of environmental

impact reviews of consumer products. In Proceedings of the 41st Hawaii International Conference on

System Sciences, 2008.

[15] R. Wash, L. Hemphill, and P. Resnick. Design decisions in the RideNow project. In Proceedings of the

2005 International ACM SIGGROUP Conference on Supporting Group Work, pages 132–135, 2005.

[16] C. Wisneski, H. Ishii, A. Dahley, M. Gorbet, S. Brave, B. Ullmer, and P. Yarin. Ambient displays: Turning

architectural space into an interface between people and digital information. In Proceedings of the First

International Workshop on Cooperative Buildings (CoBuild ’98), 1998.

Motivating Environmentally Sustainable Behavior Changes

with a Virtual Polar Bear

Tawanna Dillahunt, Geof Becker, Jennifer Mankoff and Robert Kraut

HCII, Carnegie Mellon University, Pittsburgh, PA

{tdillahu, gbecker, jmankoff, robert.kraut}@andrew.cmu.edu

Abstract

Personal choices and activities directly account for much of the energy consumption in the U.S.

with secondary impacts of those activities influencing an even larger proportion of energy

consumption. Although there is a long history of investigation into motivators for energy-

conservation, it is still unclear how to encourage persistent behavior change, and technology has

only recently been applied to the problem. In this study we sought to learn if virtual pets could

have a positive impact on real-world, environmentally responsible behavior. The polar bear is a

powerful iconic symbol for many individuals, and its fate can be simply and directly connected to

environmentally responsible behavior. Just as Tamagotchis evoked a powerful response from

their owners, we hoped to use attachment to a virtual polar bear as a motivator for energy

conservation. We ran a study exploring the impact of attachment on real-world actions. The

results of our study suggest that an interactive virtual polar bear may increase environmentally

responsible behaviors, especially when emotional attachment takes place.

1. Introduction

Global warming continues to be one of the world’s major issues. In the U.S. alone, Americans

consumed 100 quadrillion BTUs of energy in 2005 [11], with personal, individual activities

accounting for much of this consumption. For example, 40% of the energy consumed is used for

residential or commercial lighting, heating and cooling. All of this is driven directly by

individual choices or indirectly because of consumer needs. Thus, personal choices can lead to a

significant reduction in energy consumption, with a corresponding reduction in the release of

CO2, one of the primary gases responsible for climate change. However, this requires persuading

individuals to change their behavior, and maintain those changes over time, both difficult

propositions. Our approach is to leverage the power of the polar bear as a symbol of climate

change by creating a virtual pet, a technology that has promise for supporting behavior change

(e.g. [7]). We describe a pilot study showing that increased attachment to a virtual polar bear had

a significant positive impact on the number of actions that individuals had taken as reported a

week after they used our software. This work is currently being incorporated into a mobile tool

intended to influence transportation choices and into a social networking website.

There is extensive literature in the areas of environmental sociology, public policy, and more

recently, conservation psychology that discusses the promotion of environmentally responsible

behavior. In past work, we have explored the impact of motivators such as public commitment,

Figure 1: (top) a polar

bear with lots of ice

(bottom) a polar bear

with little ice

frequent feedback, and personalization on environmentally responsible behavior [10]. Research

in conservation psychology implies that animals help humans connect with nature [9]. Vining’s

literature review demonstrates the extensive evidence for strong emotional bonds between

humans and animals but she states that the answer to whether caring about animals or the

environment leads to environmentally responsible behavior remains open [12]. Technology may

provide a mechanism for leveraging these bonds to encourage behavior change. For example,

Tamagotchis are virtual pets requiring nurturing interaction in order to be sustained. This in turn

led to an emotional attachment to these virtual pets, for example owners mourned when their pets

“died” [2]. Lin and Strub’s “Fish’n’Steps” study is an example of how an interactive computer

game using a virtual pet, in this case fish, encourage physical activity [7].

Our approach integrates conservation psychology, a study which looks into relationships

between nature and humans, with persuasive technology, the study of how computers can

leverage psychological cues to motivate and influence behavior [3]. For example, motivators of

environmentally responsible behavior are more effective when they have a direct impact on

people's needs or concerns [1]. Also, caring for real animals can inspire conservation behavior

[12]. In this paper we show that an emotional connection to a virtual pet that responds to

environmental behavior can help motivate an individual by making that

behavior seem to directly impact an individual’s concerns.

2. Experiment

We conducted a one week, between subjects study to explore the effect

of higher attachment to a virtual pet (the attachment group) to lower

attachment to the same pet (the control group) on environmentally

responsible behavior. To create attachment, we used a story describing

environmental change, specifically the impact of climate change on the

habitat of polar bears, pre-tested to elicit sadness. We asked participants

in the attachment group to read it, reflect on their emotions and write

about environmental responsibility, and name the polar bear on their

display.

Participants were shown a Flash-based virtual polar bear on an ice floe

that would grow as they committed to environmentally responsible

actions and decrease as they chose not to commit to actions. As shown in Figure 1, the size of

the ice floe would change depending on the number of actions a subject committed to taking.

Our study tested the following hypotheses:

•H1: Commitments - Users who form emotional attachment to the virtual polar bear will commit

to more environmentally responsible actions than users who do not form bonds

•H2: Fulfilled Commitments - Users who form emotional attachment to the virtual polar bear

will fulfill (act on) their commitments

H1: Num Commitments H2: Num Sustained

Commitments

H3: Donations ($)

Mean Number of Units

Control

Attachment

Figure 2: (above) The attachment graph was

higher for all hypotheses, and H2 (Fulfilled

Commitments) is significant (F[1,11]=6.527,

p=.0309).

•H3: Donations - Users who form emotional attachment to the virtual polar bear will donate

more to a zoo than those from the control group

We recruited 20 subjects (10 in each

condition) associated with local universities.

Participants were given $15 and a shower

timer for their time. All participants

completed an initial survey at the start of the

experiment to test whether either group was

more likely to have higher motives for

environmentally responsible behavior and

whether this changed during the experiment.

We used a subset of questions from De

Young’s scales measuring competence and

participation [1]. Participants also completed

two additional scales both before and after

the main intervention (interacting with the

polar bear): the first measured levels of care

on agreeableness and empathy [5] and the

second measured overall environmental

concerns [6]. All scales had a reliability of .65 or higher, measured using Cronbach’s α (values

were α=.8615 for competence, α=.8920 for participation, α=.7579 for care and α=.655 for

environmental concern).

The actions are included in Table 1. Fourteen were taken from sixty actions previously designed

and tested as part of a field study of the StepGreen website [10]. We created an additional

action, “Take an environmental sustainability flyer and give it to a friend,” and made flyers

available to participants.

3. Results and Discussion

Out of the 20 participants, 60% were male and

40% female; 95% of the participants were

students. Out of the 20 participants, 19

completed the first task; we reached 11 to ask

about donations (5 control, 6 attachment), and

we were only able to reach 11 participants to ask

which actions they fulfilled (5 control, 6

attachment). The initial survey showed balance

across the groups in terms of motivation for

environmentally responsible behavior (p=.08833

and .08199).

1 Turn off the water while brushing your teeth

2 Turn off lights if you are leaving a room for more than 10 minutes

3 Wash only full loads of clothes

4 Wash only full loads of dishes

5 Dry only full loads of clothes

6 Carpool 1 day a week when you would otherwise drive

7 Pledge to ride the bus 1 day a week when you would otherwise drive

8 Combine trips in vehicle (i.e., visit multiple destinations on one trip)

9 Take the stairs instead of the elevator a minimum of 5 times per week

10 Take a shower instead of a bath

11 Restrict length of shower to 5 minutes

12 Use a low wattage night light

13 Lower heating thermostat to 68 degrees and wear warmer clothes

14 Unplug any electronic devices when not in use

15 Take an environmental sustainability flyer and give it to a friend

Table

List of Actions

As shown in Figure 2, participants in the attachment group had a higher mean number of

committed actions (H1), fulfilled commitments (H2) and a higher mean donation (H3). Of these,

the difference in fulfilled commitments was statistically significant: F[1,11]=6.527, p=.0309*.

The attachment group also demonstrated significantly greater environmental concern and greater

care after reading about and interacting with the polar bear (F[1,19]=5.1273, p=0.0369* and

F[1,19]=3.8124, p=.0675, respectively).

Our findings demonstrate that participants in the attachment group were more concerned about

the environment, and this translated directly into significantly higher reported actions. Of

particular interest is the fact that while there was not a significant difference in the number of

commitments between the groups, reported follow through in terms of fulfilling those

commitments was significantly higher. However, it is possible that participants in the attachment

group lied about their follow through out of guilt or some other emotion. Also, we do not know

how long these differences will be sustained.

4. Future Work

Our next step is to deploy the polar bear over a longer period of time, in a setting where we can

objectively measure its impact on action. This addresses the two biggest concerns mentioned

above. We plan to deploy the polar bear on a mobile platform that can track use of different

transportation options. Additionally, we are currently creating a virtual polar bear plugin for

MySpace and Facebook. The large numbers of users on sites such as MySpace (used by over

61.2 million unique visitors) and Facebook (over 19.5 million unique users)[8] presents exciting

opportunities to encourage personally- and socially-desirable change in behaviors. Users will be

able to publicly make commitments and have the status of their polar bear on display. Based on

consistency theory, users would be more likely to behave consistently with their commitment

[4]. This may also lead to further research on the impact peer pressure may have on

environmentally sustainable actions.

5. Reason for Attending

The work we described is an initial contribution to the issue of persuading individuals to take

more sustainable actions. As such, it addresses one of the three main topics of the workshop,

how to motivate sustainable action. Additionally, though there was no room to discuss it here,

our research is expanding to address the appropriate deployment platforms and messaging for

varied socio-economic and ethnic groups. We are excited by the opportunity to attend this

workshop and engage researchers on these critical fronts.

6. Author Bios

Tawanna Dillahunt is a first year Ph.D. student at Carnegie Mellon University. Her research

interests include environmental sustainability and pervasive computing.

Geof Becker works at the Tepper School of Business at Carnegie Mellon in Marketing and

Public Relations and is interested in encouraging development of online communities that

commit to positive social action.

Jennifer Mankoff is an assistant professor at CMU in the HCII. Her research interests include

environmental sustainability, pervasive computing, and Assistive Technologies.

Robert Kraut is a Herbert A. Simon Professor of HCI at CMU. He conducts research in four

major areas: online communities, everyday use of the Internet, Technology and conversation,

collaboration in small work groups, and computers in organizations.

7. References

1. De Young, R. (2003). Expanding and Evaluating Motives for Environmentally Responsible

Behavior. Journal of Social Issues, Vol. 56, No.3, pp. 509-526.

2. Donath, J. (2004). Artificial Pets: Simple Behaviors Elicit Complex Attachments. In M.

Bekoff (ed.), The Encyclopedia of Animal Behavior, Greenwood Press.

3. Fogg, B.J. (2002). Persuasive Computers Using Technology to Change What We Think and

Do. Morgan Kaufman Publishers, San Francisco, CA.

4. Harmon-Jones, E. (2002). Cognitive Dissonance Theory. Perspective on Persuasion. The

Persuasion Handbook: Developments in Theory and Practice. Sage Publications. p. 99-116.

5. International Personality Item Pool (as of February 2008): A Scientific Collaboratory for the

Development of Advanced Measures of Personality and Other Individual Differences, Big-

Five Factors Markers: http://ipip.ori.org/newBigFive5broadKey.htm.

6. Kaufman, T. (1999). A Study of the Motivations Behind Heritage Site Selection in the

United States, PhD Dissertation, Virginia Tech.

7. Lin, J., Lindtner, M.S., Strub, H. (2006). Fish'n'Steps: Encouraging Physical Activity with an

Interactive Computer Game. in Proceedings of the 8th International Conference on

Ubiquitous Computing, pp. 261-278.

8. McCarthy, C (September 13, 2007). Nielsen/NetRatings’ August social media numbers: Not

much change, CNet News.com. http://www.news.com/8301-13577_3-9777942-36.html.

9. Myers, O.E. Jr., Saunders C.D. (2002). Animals as links toward developing caring

relationships with the natural world. In Kahn and Kelert (eds). Children and Nature, 153-

178. Cambridge, MA: MIT Press.

10. Mankoff, J, Matthews, D., Fussell, S.R., Johnson, M. (2007). Leveraging social networks to

motivate individuals to reduce their ecological footprints. In Proceedings of HICSS.

11. U.S. Department of Energy (2006). Annual Energy Review 2005. Energy Information

Administration, Washington, DC DOE/EIA-0384.

12. Vining, Joanne (2003). The Connection to Other Animals and Caring for Nature. Human

Ecology Review, Vol. 10, No. 2

Points of Persuasion:

Strategic Essentialism and Environmental Sustainability

Paul Dourish

Donald Bren School of Information and Computer Sciences

University of California, Irvine

Irvine, CA 92697-3440, USA

jpd@ics.uci.edu

http://www.ics.uci.edu/~jpd

Abstract

The environmental movement, as a political body, is an alliance of many different groups with

different agendas and goals, not always compatible. Nonetheless, as a matter of political

pragmatism, the identification of “the environment” as a common focus binds these actors

together – an act of strategic essentialism. Drawing on this observation and on an account of

identity production in social software, I suggest that an alternative approach to pervasive

persuasion might be to help crystallize this sort of political connection.

1. Strategic Essentialism and Sustainability

Postcolonial scholar Gayatri Spivak (1987) coined the term “strategic essentialism” to refer to

the ways in which subordinate or marginalized social groups may temporarily put aside local

differences in order to forge a sense of collective identity through which they band together in

political movements. Post-war resistance movements to colonial rule often relied on just such

mechanisms by which particular forms of ethnicity or nation-hood were used to align disparate

groups towards common goals. Spivak’s observation is that, while such terms as “indigenous”

peoples or similar labels result in problematic and unstable groupings that erase significant

differences and distinctions (rethinking colonial categories), nonetheless these acts of identity

formation support important political ends. So while terms such as “Indian,” “African”, or

“Native American” may be manufactured and suppress highly significant differences, they

nonetheless do important work.

Studies of the social history of environmentalism show much the same sort of process at work in

the development of the environmental movement. Cronon (1995) documents the history of our

conception of Wilderness. In the period of the Western Expansion, wilderness is a threat to

human existence. It betokens the kinds of arid, unforgiving and hostile environment that settlers

might experience in their movement west, something that must be conquered (and so also a

source of opportunity.) It is not for nothing, he argues, that it is the wilderness where Christ

struggles with the Devil, or into which Adam and Eve are cast. By the late nineteenth century,

though, a new notion of wilderness has emerged – not a threat but a comfort, not something to be

overcome but something to be cherished, a place not of danger but of rejuvenation. The national

park movement reflects a change in the understanding of what wilderness might be, what it

might be worth, and why. Cronon documents a range of considerations that are part of this

ideological reframing of wilderness (including the gender issues associated with the image of the

rugged masculinity involved in taming “virgin” nature, the problems of habitation by indigenous

peoples, and the issues of the supernatural associated with the encounter with wilderness), but his

central concern is the way that the ideological construction of wilderness obscures the central

role of human action:

By imagining that our true home is in the wilderness, we forgive ourselves the homes we

actually inhabit. In its flight from history, in its siren song of escape, in its reproduction

of the dangerous dualism that sets human beings outside of nature—in all of these ways,

wilderness poses a serious threat to responsible environmentalism at the end of the

twentieth century. (Cronon, 1995:81).

More broadly, as Proctor (1998) also demonstrates, strategic essentialism lies at the heart of the

creation of the environmental movement at all. What we think of as environmentalism is a

political force resulting from the forging of an alliance between groups with concerns as diverse

as open access, biodiversity, air and water pollution, surfing, animal husbandry, agricultural

efficiency, bioengineering, and rock climbing. “The environment” emerges as a concept shaped

by the union of common interests, even though these interests might be mobilized in quite

different ways and for quite different reasons. As Spivak would note, the inherent heterogeneity

of the group is made subservient to strategic goals. Arguably, one of the reasons that the clash of

perspectives over environmentalism – between, for instance, Western environmentalists and

native Amazonians over sustainability and economic survival in the rainforest (Tsing, 2004) –

are so troubling is because they threaten the unstable alliances out of which these political

movements are formed.

2. Pervasive Persuasion

When we talk of persuasion as a consideration for information technologies, we are frequently

concerned with how behavior modification can be induced by intervening in moments of local

decision-making and by providing people with new rewards and new motivations for desirable

behaviors (Fogg, 2003). These kinds of strategies have been common, for instance, in health and

fitness applications (e.g. Lin et al., 2006). One might imagine a range of pervasive persuasive

technologies along these lines. If we argue that the essential importance of pervasiveness is that

the technology accompanies people in the course of their everyday lives, then technologies that

help people to assess their everyday actions in terms of broader questions of sustainability

present themselves – applications that help people understand their carbon footprint, for

example, or ones that provide access to environmental information as a part of shopping (e.g.

Bernheim-Brush et al., 2004).

However, if we think about environmental sustainability from a political perspective, and

particularly bearing in mind the important role of strategic essentialism, then a different

application area presents itself. From this perspective, what we might want to persuade people of

is the ways in which their interests are aligned with those of others. As is demonstrated by

sociological research into the formation of social movements, this process of alignment and

mobilization, by which one can start to find one’s own interests as being congruent with those of

others, is a critical first step in political mobilization (Snow et al., 1986).

Arguably, we can find the foundations for such technologies in the current crop of so-called

“social software” applications, of which Facebook is perhaps currently the most prominent.

Social networking sites claim simply to articulate social networks that are already there, but of

course, social networks in the formal sense are an analytic device rather than an aspect of our

own everyday experience. (This can most easily be demonstrated by asking the question, “Tell

me about someone two hops away from you in your social network.” Nobody can, since, by

definition, such a person is not known to the subject.) Instead, then, we can approach social

networking sites as technologies of affiliation, alignment, and identification, sites at which forms

of collective identity are forged and enacted.

If social software works by, first, tying individuals and actions to groups and networks and,

second, by providing a platform through which one acts as a member of a group (be that an

institutional affiliation, an informal group, or simply a identifiable social type), then a similar

approach can perhaps be harnessed in the domain of sustainability. This would suggest that,

rather than using technology to provoke reflection on environmental impact of individual

actions, we might use it instead to show how particular actions or concerns link one into a

broader coalition of concerned citizens, social groups, and organizations. So, for instance, if we

were to combine the sorts of monitoring technologies developed by Paulos (2008) or the kinds of

scanning technologies explored by Bernheim-Brush et al. (2004) with social networking

accounts of the different interests associated with sustainability debates and movements, then we

might have a system that could tell people, “the action you are about to take aligns you with X

but against Y,” or, “the products that you are looking at have these different impacts on these

different groups.” What is being done here is a process of frame bridging (Snow et al., 1986) that

not only allows for forms of reflection and behaviour modification but also links the individual

into a broader coalition of interests. What becomes visible is not so much the world, but its

political alignments.

3. Conclusions

Various attempts have been made to use pervasive technologies to connect people to the

environment in which they live (e.g. Paulos, 2008), or to provide them with tools for reflection

on the impacts of their practices (e.g. Bernheim-Brush et al, 2004). Observing that we need to

think about sustainability also as a process of political mobilization, I have suggested that we can

draw too on a different set of technologies – the social networking technologies familiar from

Web 2.0 applications – as a means to connect people not only to environments and to actions but

to other people; and, moreover, to do this in a way that helps them to see their way through the

central questions around which political action might be initiated. The new goal of pervasive

technology here, then, is to promote a form of strategic essentialism that is part of this process.

4. Acknowledgements

This work was supported in part by the National Science Foundation under awards 0205724,

0527729, 0524033, and 0712890.

5. References

Bernheim-Brush, A.J., Combs-Turner, T., Smith, M., and Gupta, N. 2004. Scanning Objects in

the Wild: Assessing an Object-Triggered Information System. Proc. Intl Conf. Ubiquitous

Computing Ubicomp 2004 (Nottingham, England).

Cronon, W. 1995. Uncommon Ground: Rethinking the Human Place in Nature. Norton.

Fogg, B.J. 2003. Persuasive Technology: Using Computers to Change What We Think and Do.

Morgan Kaufman.

Lin, J., Mamykina, L., Lindtner, S., Delajoux, G., and Strub, H. 2006. Fish’n’Steps: Encouraging

Physical Activity with an Interactive Computer Game. Proc. Intl. Conf. Ubiquitous Computing

Ubicomp 2006 (Orange County, CA), 261-278.

Paulos, E. 2008. Citizen Science: Enabling Participatory Urbanism. In Foth, M. (ed), Urban

Informatics: Community Integration and Implementation.

Proctor, J. 1998. The Social Construction of Nature: Relativist Accusations, Pragmatist and

Critical Realist Responses. Annals of the Association of American Geographers, 88(3), 352-376.

Snow, D., Ruchford, B., Worden, S., and Benford, R. 1986. Frame Alignment Processes,

Micromobilization, and Movement Participation. American Sociological Review, 51, 464-481.

Spivak, G. 1987. In Other Worlds: Essays in Cultural Politics. Taylor and Francis.

Stoler, A. 1989. Rethinking Colonial Categories: European Communities and the Boundaries of

Rule. Comparative Studies in Society and History, 31, 134-161.

Tsing, A. 2004. Friction: An Ethnography of Global Connection. Princeton, NJ: Princeton

University Press.

Bio Statement: Paul Dourish is a Professor of Informatics in the Donald Bren School of

Information and Computer Sciences at UC Irvine, with courtesy appointments in Computer

Science and Anthropology. He teaches in the Informatics program and in the interdisciplinary

graduate program in Arts Computation and Engineering (ACE). In 2008, he was elected to the

CHI Academy.

Motivation for attending: I’d like to understand the ways in which environmental questions act as

a nexus for technological, social, and cultural considerations, and are likely to develop over the

next few years.

Understanding motivation and enabling action towards

change

Penny Hagen, Duncan Underwood1

Abstract

Social Technologies can be platforms for change as they facilitate actions and act as spaces

for conversation and the dissemination of information. In this workshop paper we explore the

concept of motivation in relation to our practice as technology designers. We are specifically

interested in what motivates people to take action, and how we can support that as technology

designers.

We frame our conversation about motivation, technology, and action towards sustainability

by exploring two aspects of our design practice. The first aspect focuses on design research

and using participatory methods such as Mobile Diaries to understand what motivates people

to be involved, to take action, to contribute. The second aspect is the design of technologies,

channels or tools that enable people to act upon their motivations effectively. We provide

examples of this through case studies of our work.

1. Digital Eskimo and our stakeholders

Digital Eskimo is a design agency with a commitment to working on projects we consider to

be progressing humanity towards a nurturing (more than sustainable) way of being. To

achieve this we practice ‘Considered Design’. This definition covers a number of aspects of

how we approach our design and underlying company philosophy.

It means we are ethics and values driven. The act of doing design is meaningful to us, over

and above the notion of monetary profit. Our practice is grounded in, and informed by the

principles of sustainability. We think that change is produced collectively, created through

action, experience and sharing knowledge. And we think design can make a difference.

Our clients are equally invested in social change and are often activists themselves, existing to

promote or facilitate environmental or social causes. Many of our clients are non-government

organisations and workers unions. Similarly the stakeholders, users or community members

that we are designing for are also (often) interested in change - seeking action or ways to

complete action. Our work talks to their motivations and interests. Some of our work focuses

on creating and growing existing communities over time, some of our work focuses on

1 Digital Eskimo, Level 4, 16 Foster Street, Surry Hills NSW Australia

www.digitaleskimo.net

ph@digitaleskimo.net

capturing focused action one time, on a large scale. Our work is underwritten by an approach

to design that focuses on understanding what motivates people to take action or be part of a

community towards change.

While our core work is manifest in online technologies, it increasingly takes the form of more

integrated services across multiple delivery systems, overlapping with ubiquitous and

pervasive systems. Through the workshop we hope to explore how our approach to design can

be applicable to pervasive systems, and contribute to expanding discourse on how we as

designers we can build pervasive technologies that support or motivate sustainable living

practices.

2. An approach to design

Our work is guided by the traditions of Participatory Design [2][7], in that we recognize the

expertise of the users and stakeholders, what Sanders calls a Participatory Mindset [6]. For

these reasons we select research methods based on their ability to immerse the design team in

the world of our stakeholders, methods that facilitate a collaborative relationship with clients

and aim to enable all stakeholders to be appropriately represented in the design. We use

participatory methods such as workshops and collaborative brainstorming and we build

collective personas and scenarios with our stakeholders [3]. While we continue to employ a

range of researcher driven methods such as interviews, focus groups, surveys and

questionnaires we embrace methods that are more designerly (e.g. [1]) participatory (e.g. [8])

and interventionist (e.g. [5]) such as Mobile Diaries [4].

In this paper we briefly present three diverse “social technology” case studies where actions

or discourse are facilitated through technology. The first is a Design Research case study

using technology to understand underlying attitudes to sustainability as well as ways to

support corresponding actions. The second and third, an Integrated Campaign and an Online

Campaign site respectively both aim to motivate and facilitate change by using technology to

amplify individual actions on a collective platform, albeit in quite different ways.

3. Case studies

3.1 WWF Human Habitat Diaries (Design Research Project)

Digital Eskimo conducted a research project with five participants over a two week period.

Each was selected from one of the different demographic groups likely to interact with

WWF's FutureMakers project. The intention was to provide a window into the lives of

participants, a way for them to share their world with us, and their perspective on it. This

research was to inform the design of online tools that enabled and encouraged such sharing

between people on an ongoing basis.

Participants were prompted to reflect their environment and daily lives from a sustainability

perspective and were invited to implement one of eight measures that would reduce their

negative impacts on the environment. Participants used Mobile Diaries which include camera

phones, low resolution video cameras, maps and notebooks to record their daily experiences

at work and at home. SMS’s and images with audio annotations were sent from the phone to

personal Habitat Diaries (password protected blogsites) assigned to each participant that could

be accessed by the participants and the designers/researchers. The notebooks, video cameras

and maps were mailed back to us at the conclusion of the study. Combined, they enabled us to

develop a rich picture of the participants’ lives, habits, and attitudes to sustainability. The data

gathered enabled us to understand the participants motivations around issues of sustainability,

the obstacles motivated people faced in making change, as well as identify when and why

they might be motivated to participate in an online community focused on supporting change.

3.2 David Hicks – Amnesty International Australia (Design Case Study)

In this case study Digital Eskimo was asked to develop an online facility to promote a project

of Amnesty International Australia (AIA), raising awareness of the issue of David Hicks’

imprisonment without trial in Guantanamo Bay.

AIA had built a full sized replica of the cell David Hicks lived in and the public were invited

to spend a small amount of time in the cell alone to consider David’s experience in

Guantanamo Bay and the broader issue of his ongoing imprisonment. We saw an opportunity

to use the online space to show and share people’s reaction to their time in the cell as well as

their attitude on the issue by placing a webcam in the cell that enabled visitors to record a 30

second message to camera that was then syndicated to one or more websites. The

corresponding website enabled people who could not access the cell to understand a little of

the experience, at the same time by broadcasting these messages we affected a far broader

audience than would have originally have been reached.

Digital Eskimo’s response to this opportunity promoted the tour of the cell and communicated

AIA’s campaign objectives. But furthermore it allowed people to act upon their motivation to

do something about these issues: voice their opinion on the issues of David Hicks’ detention

and the imprisonment of terror suspects without trial; to hear the opinions and responses of

others to their experience in the cell; and to share their own views among their networks.

3.3 ACTU Your Rights at Work campaign site

The website is the hub of the ACTU’s 150000 member campaign against the unpopular

WorkChoices legislation. Our team designed the site with an emphasis on informing workers

and inspiring specific targeted action.

The site delivers the ACTU’s information in a simple and effective style, as well as telling the

stories of workers who have been affected by the WorkChoices legislation. Primarily however

the site is a campaigning tool that was designed to deliver single strong calls to action that can

be updated by the ACTU at a moments notice.

This tool allowed the ACTU to easily and quickly communicate to their member base. It

enabled motivated people to undertake a collective action and to easily promote such action to

their friends and networks, amplifying the call to act and directing such activity in order to

maximize its effect.

4.0 Conclusion

Digital Eskimo is a design agency that utilizes social technologies to promote and progress a

nurturing (more than sustainable) way of life. Key to our own motivation within our projects

is that we share the goals of our clients. Key to achieving successful design outcomes is

developing a richer understanding of the motivations of each project’s stakeholders and our

ability to design technologies that encourage and enable people to act upon these motivations;

connecting, sharing knowledge and working together to create change towards a nurturing /

sustainable life.

5.0 Biography

5.1 Penny Hagen

Penny Hagen is the Executive Producer at Digital Eskimo. Prior to joining Digital Eskimo

Penny spent 10 years freelancing as a producer, designer and trainer specializing in

interactive media and community projects in Sydney and New Zealand. Penny has also

worked as a researcher for UTS’s Interaction Design Lab investigating mobile technology use

and conducting research into social software and participatory design methods. Underlying

her approach to technology design is a commitment to creating communication tools that

acknowledge the emergent nature of social networks and encouraging appropriation by the

people that use them.

5.2 Duncan Underwood

Duncan Underwood is the Sustainable Development Manager at Digital Eskimo. Duncan

graduated with a Bachelor of Design (Industrial) from UTS in 1997 and is currently enrolled

in Master of Social Science (International Urban and Environmental Management) at RMIT.

Before Digital Eskimo, Duncan worked in business development at Clean Up Australia.

6.0 Bibliography

[1] GAVER, B., DUNNE, T. & PACENTI, E. (1999) Design: Cultural Probes.

Interactions, 21-29

[2] GREENBAUM, J. & KYNG, M. (Eds.) (1991) Design at Work: Cooperative Design

of Computer Systems, New Jersey, Lawrence Erlbaum Associated.

[3] HAGEN, P. & GRAVINA, D., Sharing through Artefacts: client-user centred design.

Ozchi. Sydney, ACM, 2006

[4] HAGEN, P., ROBERTSON, T. & GRAVINA, D. (2007) Engaging with stakeholders:

Mobile Diaries for social design. DUX. Chicago, USA, ACM.

[5] PAULOS, E. and JENKINS, T. Urban probes: encountering our emerging

urban atmospheres. ACM Press, Portland, Oregon, USA, 2005

[6] SANDERS, E. B.-N. (2006). Design Research in 2006. Design Research Quarterly, 1.

[7] SCHULER, D. & NAMIOKA, A. (Eds.) Participatory Design: Principles and

Practices. NJ, Lawrence Erlbaum Associates.

[8] VISSER, F. S., STAPPERS, P. J., LUGT, R. V. D. & SANDERS, E. B.-N. (2005)

Contextmapping: experiences from practice. CoDesign, 1, 119-140.

THE NEW WELL-TEMPERED ENVIRONMENT: TUNING

BUILDINGS AND CITIES

Dan Hill & Duncan Wilson

Abstract: Our work has two main components - one part practical, the other theoretical - though

both are related, and have developed in discussion. They both concern the measurement and

feedback of energy usage in buildings, and then cities, in order to enable users to change their

behaviour - to in effect, ‘tune’ their environment. The work is intended to raise important questions

about the efficacy of such schemes, outline the potential of systems when stretched over urban and

social networks, to speculate about the future of information as part of the fabric of buildings, and

to make some practical considerations clear along the way.

1: Tuning 13 Fitzroy Street

The first part of the work concerns research and development work led by Arup’s Dr. Duncan

Wilson, deploying wireless sensor networks in the new Arup building at 13 Fitzroy Street, London.

This is predicated on the emerging understanding that occupation of a commercial building

typically costs more than 200 times the initial capital cost of construction and that ~40% of final

energy consumption in the European Community is in the buildings sector. In addition governments

are pushing regulatory frameworks to reduce the energy impact of such running costs [1]. Within

the UK, energy used in buildings accounts for some 50 per cent of the country’s carbon dioxide

emissions.

Thus the ability to moderate the use of energy in buildings can have a significant impact on the

overall efficiency of the built stock. Changes to the design and construction of new buildings is

leading to significant improvements in building efficiency but the majority of the built stock is

comprised of older buildings with limited potential for passive energy saving. The Living Buildings

initiative [2] at Arup is one example of our sustainability activities working towards the

consequences of the context above. At a macro level we are also working at a city scale on projects

such as Dongtan eco-city in China [3]. But what are the opportunities for pervasive computing?

We have been exploring the use of ubiquitous or pervasive computing technology based around

wireless sensor network (WSN) platforms to gain improved understanding of real-time operation of

buildings and to allow greater environmental control of existing buildings that do not have the

installed infrastructure to actively control energy consumption.

The first small steps have been to deploy WSN's in the office environment to investigate both

energy consumption of the building and the comfort of the space [4]. Various quantitative and

qualitative measures are being taken and fed back in real-time to occupants of the space (not just

building / facility managers). At the newly built Fitzroy Street building, the two main sensor

network development platforms used were Crossbow and Arduino. Motes have been deployed

throughout the building (and in a related study at Central Saint Martins college in London),

connecting over Bluetooth IEEE 802.15 and sensing occupancy levels (via Passive InfraRed

sensors) combined with performance data from the building management system. The Bricks

Framework is used to fuse data together, and provide the base for dynamic visualisations, sound

installations and interactive devices.

We’ll discuss various methods for conveying information back to users, and the many issues

therein. As it stands, the system is producing data across all axes and various visualisation methods

are being designed for a number of locations within the building.

Ultimately, this work at the level of buildings may provide an opportunity to start doing city wide

monitoring of the pre- and post-occupancy performance of sustainable buildings. This pervasive

data collection from very large sensor populations could be integrated to support control and

optimisation at this scale. City information modelling systems are beginning to emerge,

extrapolating from Arup’s work with building informational modelling (BIM), and this provides a

unique opportunity for post construction analysis and validation.

2: The Personal Well-Tempered Environment

Developed by Dan Hill, this section continues the themes covered thus far, extrapolating the

potential for such schemes when stretched across urban and social networks. It’s an imagined

system at this point [5], a real-time dashboard for buildings, neighbourhoods, and the city, focused

on conveying the energy flow in and out of spaces, centred around the behaviour of individuals and

groups within buildings.

In this, it becomes a form of 'BIM 2.0' that gives users of buildings both the real-time and

longitudinal information they need to help change their behaviour and thus use buildings, and

energy, more effectively. It would be an ongoing post-occupancy evaluation for the building, the

neighbourhood and the city. Importantly, it proposes measuring contribution as well as

consumption, through sensors embedded into localised wind and solar power, grey-water collection,

and so on. Further, it explores the idea of measuring behaviour across wider circuits, such as an

individual’s movement through the city (scoring ‘points’ for public transport versus private

transport, monitoring environmental usage in office environments, and so on.)

Suggested as a software service layer for connecting things together within and across buildings, it

would take a ‘plug-in’ approach to connecting energy sources and resources, drawing from

architectural theory of Archigram and Cedric Price amongst others. Multi-sensory feedback is a

particular theme in the suggested interface, exploring different ways of conveying this information.

The work is based on a survey of existing energy monitoring schemes and products in this area,

collated and discussed online. Over and above this, it folds in some ideas from social software,

particularly the reflexive mode produced by systems such as Last FM, Flickr, Dopplr, Nike+ and so

on, and wraps these up with the aforementioned architectural theory as well as making connections

to newer concepts like Bruce Sterling’s spimes and everyday product design.

Drawing from this recent history of social software, the proposal describes various ways in which

an ‘open’ approach to data, allied with social networks, may enable a socialising of the data, or

even a ‘gaming’ element - in which individuals, suburbs, neighbourhoods and cities can compare

their environmental performance. This latter aspect is an attempt to make the civic relationship

between an individual and their environment clear, thus addressing a key issue in the emerging

informational city. Taking the conceptual starting point of an API on a house, the idea suggests

extending this to the API on the neighbourhood, even the city itself. Making the effects of

informational behaviour visible on the street in turn asks further questions about how to perceive

and communicate the emerging informational aspects of the contemporary city, a theme

increasingly fundamental to urban planning.

Though initial investigations elsewhere suggest that feedback on energy use can help change

behaviour [6], questions will be asked of the efficacy of such ‘persuasive visualisation’, as part of a

critical assessment of whether such systems can truly have a beneficial effect in terms of ‘tuning the

environment’, or whether the real problems lie elsewhere.

The connection to the aforementioned Arup projects will be made clear, seeing the Personal Well-

Tempered Environment as an example of Arup’s approach of ‘total design’ - a multidisciplinary

framework for building - and discussing how information itself can increasingly be thought of a

material within building. With a holistic approach, it makes sense to consider information as part of

the built fabric, just as with glass, steel, ETFE etc. The Well-Tempered ideas will begin to inform

Arup’s work on the ground, and vice versa, thus creating a constructive relationship between

imagined informational architectures and pragmatic, deployed engineering.

Motivation for participating in the workshop

This workshop provides an opportunity to extend the thinking around our projects so far, testing the

ideas against related work and discussing with others working in this field, subsequently wrapping

the findings back into practical work through our projects and research at Arup.

Biographies

As part of the Foresight Innovation and Incubation team at Arup, Dr. Duncan Wilson (duncan.wilson@arup.com) is

responsible for researching medium and long term futures with a focus on social and technology factors. He develops

foresight and innovation capability within Arup, co-created the Drivers of Change concept, and programme manages a

series of workshops on the future of the built environment. He was Principal Investigator on a Euro 1.4 million two year

research project (DTI technology programme) applying wireless sensor networks in the built environment, is a partner

in the European Union SENSEI project looking at networks of wireless sensor networks and is leading an internal

research project on the implications of ubiquitous computing for Arup. His research merges interests in sensing and

monitoring and creating interactive, ambient displays that solicit and feedback information with the intent of

influencing behaviour. Duncan is a Chartered Engineer (IET), has a PhD from University College London in Artificial

Intelligence and Machine Vision and blogs at http://www.driversofchange.com/emtech/

NB: The work at Arup is supported by two UK national programmes (DTI/TSB Technology Programme)

Dan Hill (cityofsound@gmail.com) has been working at the forefront of innovative information technology since the

early '90s, and is responsible for many innovative, popular and critically acclaimed products and services He was Head

of Interactive Technology & Design at the BBC in London for 5 years, before launching the critically-acclaimed

international magazine Monocle during 2007, responsible for its digital services. He recently joined Arup as a Senior

Consultant in their planning group, working with urban informatics. During 2007, Hill co-organised the Postopolis!

architecture and urbanism exhibition/conference in New York City, and has a background in academic research and

teaching in urban regeneration and urban informatics. His weblog City of Sound (http://www.cityofsound.com/) is

generally considered to be amongst the foremost architecture and urbanism sites, recently voted by Planetizen as one of

the ten best planning, design, and development sites for 2008 (http://www.planetizen.com/websites/2008).

References:

[1] For example, the EU Energy Performance of Buildings Directive, http://www.buildingsplatform.org/cms/

[2] http://www.arup.com/advancedtechnologyandresearch/feature.cfm?pageid=9887

[3] http://www.arup.com/eastasia/project.cfm?pageid=7047

[4] FANGER, P.O. Thermal Comfort. Analysis and applications in environmental engineering. McGraw Hill.

http://www.flickr.com/photos/pseudonomad/collections/72157603889824626/

[5] http://www.cityofsound.com/blog/2008/01/the-personal-we.html, with video at

http://nextbutton.pureprofile.com/TVC/?id=4326

[6] Pacific Northwest National Laboratory of the Energy Department, 2008

(http://www.nytimes.com/2008/01/10/technology/10energy.html?scp=3&sq=department+of+energy)

PROMOTING ENVIRONMENTALLY SUSTAINABLE

BEHAVIORS USING SOCIAL MARKETING IN

EMERGING PERSUASIVE TECHNOLOGIES

Omar Khan and John Canny

Berkeley Institute of Design, Computer Science Division,

University of California, Berkeley, USA

{omar, jfc}@cs.berkeley.edu

Abstract

We argue that social marketing, a strategy that uses techniques from corporate marketing to

influence the behavior of target audiences, is a useful framework for thinking about motivating

people to enact environmentally sustainable behaviors. We critically examine some pervasive green

applications through the lens of social marketing and discuss how we might study various

persuasive factors encouraged by social marketers in these domains and in our own research.

1. Introduction

Social marketing uses tools from corporate marketing to influence the behavior of target audiences

[1, 6]. Kotler and Zaltman, reflecting on successful marketing campaigns for products like soap,

observed that some principles might translate to selling social causes. In the ensuing 35 years,

social marketing has become particularly widely accepted in the public health domain, where it has

been successfully used to influence behaviors concerning drunk driving and childhood obesity, for

example. In this paper, we suggest using principles from social marketing to enhance persuasive

technologies and help focus them on changing behaviors, and not just raising awareness.

2. Social Marketing: Setting the Stage

In his recent text [1], Andreasan describes a set of concepts, many from corporate marketing, that

make the social marketing process effective. While social marketers emphasize the importance of

considering the factors listed below, the degree of effectiveness of some of these factors has not

been carefully studied. To that extent, one important research contribution will be the study of the

effectiveness of these factors in the area of encouraging environmentally sustainable behaviors.

Benefits and Costs

Like obtaining a product, when an individual chooses to enact a certain behavior, they must pay

some cost, and hopefully acquire some benefits. The social marketer's challenge is to sell the

benefits while minimizing the costs. Related to this is how the benefits are presented. Research

shows that the order in which requests of individuals are made can have a significant effect on

engagement behavior [7]. In addition, the method of presentation can be very important [10].

Other People

Robert Cialdini has shown that 1) when deciding on a behavior to enact, people are strongly

influenced by knowing what others are doing and 2) in certain situations, people report that they do

not think they will be strongly influenced by others, but in fact these same people appear to be most

strongly influenced by what others are doing [2]. This perhaps surprising result means that social

influence can be particularly powerful because people do not guard themselves against such

influence.

Self-Assurance (or Self-Efficacy)

Even if benefits, costs and other people are aligned in favor of an individual enacting a certain

behavior, that individual may still not act. One reason is that they may believe they cannot enact the

behavior. This is where we must provide support mechanisms like support groups (e.g. Alcoholics

Anonymous) and skills training [13].

Segmentation and Identity

Because of large variability in target audiences, it is unlikely that treating the audience as one large,

coherent market will be successful, and thus we should perform market segmentation. Identity-

based marketing is related to this idea. Controlled studies have shown that if individuals with a

relevant identity (say they are "green" individuals) that is primed (the individual is given content

that surfaces "green" thoughts) are then much more likely to purchase a product related to that

identity when compared to green individuals who were not primed [11].

3. Pervasive Green Applications through the Social Marketing Lens

3.1. Highly Sensed Virtual Environments

Green social networking site applications have been discussed in the literature [8] and appear

online. One popular green application on Facebook is called "I Am Green" [4]. Users provide the

application with a list of their green behaviors. Each green behaviors gets you a leaf, and you are

compared to your other friends who have also installed the application. As a leaf collecting

competition, it may be effective, but it is unclear if it is actually effective at advocating and

motivating users to enact environmentally sustainable behaviors.

Consider the profile view of the "I Am Green" application in Figure 1. What is most prominent is

the number of leaves the friends have, not the behaviors they enact. To leverage social influence,

the application could instead say "4 of your friends recycle, even when it is not convenient." If four

of my friends do it, based on Cialdini's work, we can hypothesize that we are already more likely to

enact that behavior. Furthermore, I could click on the behavior and learn more about it, like its

benefits and costs. Similarly, popular behaviors could be advertised.

Finally, recall the social marketer's emphasis on audience segmentation. Social networking sites

provide such detailed information about individuals and their social network that creating audience

segments of size one is possible. Indeed, we hypothesize that presenting users with recommended

behaviors based on collaborative filtering instead of the most popular behaviors will lead to

increased adoption of the recommended behaviors.

Figure 1: Screenshot from the I Am Green Facebook application

3.2. Dormitory Energy Competition at Oberlin College

On the Oberlin campus in 2005, an energy saving competition was run between dormitories [9].

The dormitories that saved the most energy, over a certain period, would win a prize. Building on

the well documented effect of providing energy consumption feedback to reduce future

consumption [3], the researchers provided one group of dormitories with real-time consumption

information they could view on the Internet (see Figure 2 below for a particular residence, Kade

Hall), or on an interactive display in the lobby of the building. Dorms engaged in the competition

using these advanced monitoring systems saw significant energy reduction, over and above those

dorms that did not have such detailed monitoring technology. However, examination of energy

consumption patterns after the end of the competition suggests that the numbers have returned to

near their original, pre-competition baseline [5]. From a social marketing perspective, this is not

surprising. One of the primary benefits offered to students in the dorms was the potential to win the

competition. When the competition is discontinued, both the benefit of having a prize, and the

benefit of friendly competition, disappear, and what's left is the somewhat intangible benefit of

reducing the campus' electricity consumption and maybe indirectly helping the planet. This is one

hypothesis for the return to the baseline. We could examine the benefits hypothesis in future

competition by redesigning the benefits to be seen as continually useful. Another possibility is to

ensure the benefits remain for a long enough period so that individuals internalize their behaviors.

Such work has been done in residential settings.

4. Our Work

We intend to study the effectiveness at reducing consumption of the social marketing concepts

outlined above, and game-like mechanisms (e.g. competition, scoring points) as demonstrated in the

Dorm Energy Competition. First, we can perform basic experiments in a laboratory setting to study

the impact of different social marketing factors incorporated in technologies on behavioral change

(as in [2, 7, 10]). In the field, we hope to study these factors in two domains:

4.1 Highly Sensed Virtual Environments

In an online social networking environment, we will build an application that promotes

environmentally sustainable behaviors. The application may be similar to the "I Am Green"

application, in that users must select behaviors that they enact. Concretely, we would then

manipulate the persuasive factors mentioned above for different groups, and monitor the uptake of

behaviors. Here are some of the persuasive factors we may study:

Social Influence: since the application sits on top of a social network, we can leverage information

about friends. We can present performance information about friends, make comparisons between

individuals and their friends

Game Mechanisms (scoring, competition): behavior choices might be translated to a score (like

leaves in "I Am Green") and can be billed as a competition amongst participating individuals

4.2 Reducing Individual Energy Consumption in Office Spaces

Office buildings consume a huge proportion of energy in most countries. Lighting and electronics

usage by individuals is a significant component of office building energy consumption. We are

building a dashboard display for an office space. Based on survey work by the Center for the Built

Environment at Berkeley, often office occupants feel they have very little control over energy

consumption in their space. Thus, one experiment we may run is between visual displays that only

display consumption information, and those that display consumption information and promote

appropriate behaviors for changing consumption in the space.

Game mechanisms may also be effective in this domain. A study [12] showed that competition

between office spaces yielded increased reductions when compared to spaces that did not compete.

5. About the Authors

Omar Khan is a 3rd year PhD student studying human computer interaction. He is trying to persuade

individuals to enact environmentally sustainable behaviors. John Canny is the Paul and Stacy

Jacobs Distinguished Professor in the UC Berkeley Department of Computer Science. His research

interests include activity-oriented design and educational and persuasive information systems.

6. References

[1] ANDREASEN, A.R., Social Marketing in the 21st Century, Sage Publications Inc, 2006.

[2] CIALDINI, R.B., “Basic Social Influence Is Underestimated,” Psychological Inquiry, vol. 16, 2005, pp. 158-161.

[3] DARBY, S., The effectiveness of feedback on energy consumption. A review for DEFRA of the literature on

metering, billing, and direct displays, University of Oxford, Environmental Change Institute. April, 2006.

[4] "I Am Green," January 2008, http://apps.facebook.com/iamgreen/.

[5] Lucid Design Group (private communication), 2007.

[6] P. KOTLER and G. ZALTMAN, “Social Marketing: An Approach to Planned Social Change,” Journal of

Marketing, vol. 35, 1971, pp. 3-12.

[7] W. LIU AND J. AAKER, "The Happiness of Giving: The Time-Ask Effect," To appear in the Journal of Consumer

Research, February 2008.

[8] J. MANKOFF et al., “Leveraging Social Networks To Motivate Individuals to Reduce their Ecological Footprints,”

Proceedings of the 40th Annual Hawaii International Conference on System Sciences, 2007, p. 87.

[9] J.E. PETERSEN et al., “Dormitory residents reduce electricity consumption when exposed to real-time visual

feedback and incentives,” International Journal of Sustainability in Higher Education, vol. 8, 2007, pp. 16-33.

[10] D. RAMACHANDRAN and J. CANNY, "The persuasive power of human-machine dialogue," To appear in

Proceedings of Persuasive 2008.

[11] REED, A. “Activating the self-importance of consumer selves: Exploring identity salience effects on judgments,”

Journal of consumer research, vol. 31, 2004, pp. 286-295.

[12] F. SIERO et al., “Changing Organizational Energy Consumption Behaviour Through Comparative Feedback ,”

Journal of Environmental Psychology, vol. 16, 1996, pp. 235-246.

[13] V.J. STRECHER and M.E. DEVELLIS, “The Role of Self-Efficacy in Achieving Health Behavior Change,”

Health Education & Behavior, vol. 13, 1986, p. 73.









Seng W. Loke,

Jugdutt Singh,

and Hai Le

Abstract

We propose a technological solution to the general problem of empowering individuals to take

actions related to environment sustainability, going beyond mere reminder systems or simply

passive written guidelines. Resource-usage policies at the home, community, state or national level

will be encoded in a formal rule-based language and so, translatable to policies or goals at lower

levels of granularity.

1. Introduction

This position paper outlines a technological solution to the general problem of empowering users to

take actions related to environment sustainability, going beyond mere reminder systems or simply

passive guidelines. Our position is that such a system can make a difference. Several systems have

been proposed in the water and energy domain [1,2,5], in order to encourage prudent use of

resources based on the idea of persuasive technology [3]. More generally, the idea is to develop

systems which can autonomously

(i) quantify effects of actions and measure consequences (whenever measurable) – e.g.,

actions related to usage of resources such as water and energy, and CO2 emissions; such

quantification can be in terms of low-level activities: for example, what is the cost of

this handwash? How much of CO2 emissions will leaving this device running till I

return from dinner cost me? etc

(ii) help users be aware of such effects and consequences, and then

(iii) facilitate users adjusting their behaviour or attitude with respect to goals related to these

effects and consequences.

The goals related to resource usage or CO2 emissions may be specified as policies at different

levels, by users themselves, by the local council, the state government or even a national body. User

goals might be shaped by policies from above. A system which can automatically map a high level

national goal to goals tailored for individuals would help individuals make a difference.

Department of Computer Science and Computer Engineering, La Trobe University, Melbourne, Australia

(s.loke@latrobe.edu.au)

Center for Technology Infusion, La Trobe University, Melbourne, Australia (Jack.Singh@latrobe.edu.au)

Center for Technology Infusion, La Trobe University, Melbourne, Australia (hai.le@latrobe.edu.au)

2. Architecture

Buildings blocks of technology is available for a system automating this process of taking high level

goals and helping users “digest” and be influenced by these goals in daily life. We outline the key

aspects and associated technologies of such a system, as follows:

• Fine-grained metering and monitoring: ideally, devices should be available to capture

usage of resources at different outlets (e.g., for water, be able to monitor usage at different

sinks or faucets) as well as usage of the entire home or floor or building, and perhaps, with

some instrumentation, resource usage of particular individuals in the home.

The problem of labeling an instance of resource consumption (which we define as any

distinguishable action or activity employing a countable measure of resources, often, but not

necessarily, having an identifiable start and end time, e.g., opening a kitchen water faucet for

a certain period, or switching on a lamp for certain time) with its particular time, place, actor

(the user of the resource), and purpose is generally difficult though not impossible with

adequate instrumentation for the user. Traditional metering is, hence, merely confined to

units of resources used but more information about how and why resources are used will

help computer automation of resource-control, as we discuss further below.

• Data processing and situation understanding: once resource usage can be tracked, one

would need to process the data from such metering, either to translate processed data into

appropriate visualization forms and various status displays to simply inform users, or more

elaborately, to trigger particular persuasion strategies to influence users’ (or consumers’)

behaviours and attitudes towards goals (which may encompass status displays, but more

than that, also other persuasive messages or actions to take (e.g., reducing water flow in a

long shower) according to persuasive techniques being employed), as we consider further

below.

The system could also perform longer term analysis of metered data to determine usage

trends over days, weeks, months or even years. Understanding situations of use of particular

units of resources can help inform the system about what actions to take or messages to use

at that instant. There has been tremendous amount of work in sensor-based inference of

users’ context and situations [4], as well as inferring user’s current activities.

Useful

context here (with regards to water, say) include the identity of individuals, the location

where water is being used, the time in which water is being used, the activity for which the

water is being used and the urgency of the use, current water costs, user-specified cost/water

usage goals, current water levels and current policies on water restrictions, all of which

aggregates into situations of use, which can then be mapped to appropriate persuasive

strategies and messages.

For a given type of resource, models of what constitutes normal resource usage, wasteful

usage, and conservative usage will be needed. Further finer demarcations than these three

might be useful, or fuzzy categories.

• Action strategizing: Given an instance of resource consumption, a system will have rules

which could map the collection of (i) metered data, (ii) usage trend knowledge, (iii)

http://www.activity-based-computing.org/

computed effects of the resource consumed, (iv) policies (at home, community, state or

national level) and associated goals, and (v) the inferred situations of use, to actions to

regulate usage in that instance (possibly even identifying wastage or non-usage).

Actions can range from simply notifying users, i.e. displaying to users cost or water levels in

a visual form, advice on water-saving for specific tasks, various forms of reinforcement

messages, just-in-time prompts, social validation (e.g., where possible show the best water

users in the home), adaptations (according to usage history or current needs), negotiation

(e.g., to keep to a previously specified budget, the user can use more water this time but

have less to use next time), recommendations of water saving devices, to taking action on

behalf of the user (e.g., stopping water flow at certain times – if the user so authorises such

pre-settings).

• Feedback and strategy revision: there is a cycle of monitoring resource usage and

situations of use, adopting a course of action and a corresponding persuasion strategy,

following the strategy, and then adjusting or revising strategies midway depending on

detected changes in resource usage (e.g., due to users’ behavioural change). Such a cycle of

processing is akin to the paradigm of knowledge-based intelligent agents [6], which runs in

the “background”, as depicted in Figure 1.

Figure 1: Overall system behaviour – a cycle of sensing/monitoring resource usage, reasoning about the way the

resource is used and acting

3. Conclusion

The above system is under development and we plan to build a shell in the expert systems style

which can be instantiated with different models of resource usage behaviour, persuasion strategies,

action modules, and knowledge about what actions to take in different situations, to help manage

the usage of a particular type of resource. Such a system can be interfaced to various resource

monitoring devices and displays for persuasive messages.

Resource-usage policies at the home, community, state or national level will be encoded in a formal

rule-based language and so, translatable to policies or goals at lower levels. For example, a goal for

water usage at the home level can be created based on a community policy. The system will then,

while monitoring resource usage or computing its effects (e.g., greenhouse gas emissions), adopt

various persuasion strategies and interact with the user to help meet this goal.

Thereafter, experiments with the system and usability evaluation in real settings will be done with

users. There are issues related to producing and employing the necessary monitoring equipment for

the systems we propose here, since that itself could lead to further energy consumption (which

remains to be measured), and raise concerns about privacy. It is possible to only use monitored

information for the purposes of providing advice to the user or prudent messages, but this implies

careful safeguarding of gathered context information (indeed mechanisms to allow users themselves

to regulate context information or protect privacy has been considered elsewhere, e.g., [7,8]) – not

insurmountable but possible with existing policy-based solutions. In addition, while CO2 emissions

might not be decisively quantified at this time, relative measures might be applicable and usable in

our approach. Lastly, our solution relies on prudent persuasion rather than coercion, and so, it is

possible for individuals to ignore the messages of our system – the role of our system is, hence, to

empower, encourage, and facilitate those already desiring to make some difference. While this

paper has proposed a technological solution, further social and cultural implications of our proposal

remains to be explored.

Acknowledgements. We thank anonymous referees for their valuable comments on this paper.

4. References

[1] MCCALLEY, T., KAISER, F., MIDDEN, C.J.H., KESER, M., AND TEUNISSEN, M. Persuasive appliances: Goal

priming and behavioural response to product-integrated energy feedback, PERSUASIVE 2006. LNCS 3962, New York,

Springer Verlag, pp. 45-49, 2006.

[2] BONANNI, L., ARROYO, E., LEE, C.-H., AND SELKER, T. Smart Sinks: Real-World Opportunities for Context-

Aware Interaction, in CHI 2005, pp. 1232-1235.

[3] FOGG, B.J. Persuasive Technology: Using Computers to Change What We Think and Do, Morgan Kaufman

Publishers, 2003.

[4] LOKE, S.W. Context-Aware Pervasive Systems: Architectures for a New Breed of Applications,"Auerbach

Publications (CRC Press), 2007.

[5] AL MAHMUD, A., DADLANI, P., MUBIN, O., SHAHID, S., MIDDEN, C.J.H., AND MORAN, O. iParrot:

Towards Designing a Persuasive Agent for Energy Conservation, in de Kort et al. (eds), PERSUASIVE 2007, LNCS

4744, New York, Springer Verlay, pp. 64-67, 2007.

[6] RUSSELL, S.J. AND NORVIG, P. Artificial Intelligence: a Modern Approach: Prentice-Hall, 1995.

[7] HULL, R., KUMAR, B., LIEUWEN, D., PATEL-SCHNEIDER, P., SAHUGUET, A., VARADARAJAN, S., AND

AVINASH VYAS, A. Enabling Context-Aware and Privacy-Conscious User Data Sharing. Proceedings of the IEEE

Intl. Conf. on Mobile Data Management (MDM 2004), 2004. http://db.bell-labs.com/project/e-services-

customization/privacy-conscious-data-mdm2004.pdf

[8] HONG, J.I. An Architecture for Privacy-Sensitive Ubiquitous Computing. PhD Thesis, University of California at

Berkeley, Computer Science Division, Berkeley, 2005. http://www.cs.cmu.edu/~jasonh/publications/jihdiss.pdf

Biography

Dr. Seng Loke is a Senior Lecturer at the Department of Computer Science and Computer Engineering, La Trobe

University, Melbourne, Australia. He has authored and co-authored more than 170 research publications, and leads the

Pervasive Computing Group at La Trobe. Prof. J. Singh is Director of the Center for Technology Infusion, at La Trobe

University, Melbourne, Australia, and Research Professor. Dr. Hai Le is Research Fellow at the Center for Technology

Infusion, at La Trobe University. The team, recently formed, works on cost-effective computing and engineering

solutions to the important issues of environmental sustainability.

The workshop will be highly useful as a forum for interacting with experts in other fields, and in creating a

multidisciplinary understanding of the issues (and perhaps solutions) at hand.

Using persuasive technology to encourage sustainable behavior

Cees Midden, Teddy McCalley, Jaap Ham & Ruud Zaalberg

Eindhoven University of Technology

The Netherlands

c.j.h.midden@tue.nl

Abstract

In this paper sustainable consumption is conceptualized as the result of various types of interactions between

users and systems. We review attempts to promote sustainable behavior and discuss contributions by using

persuasive technology. In particular, we focus on the appraisal of climate risks and interactive approaches to

influence energy consumption in households.

1. Introduction

The impact of human activity on the natural environment has severely affected the ecosystems on earth and in

the long run might lead to serious threats to human life and civilization. The environmental impact of humans

can be roughly assessed as a function of their numbers, their affluence, and the technology they currently use

(cf. Ehrlich & Ehrlich, 1991). However, despite the fact that humans have used technology as long as they

have consumed natural resources, technology as related to environmental resource use, is often set apart from

the study of human behavior and resource conservation. This separation has hampered interventions to protect

natural resources and constrain negative environmental impacts.

Various studies have shown that a purely technological approach to reduce energy consumption often leads to

disappointing results due to changes in user behaviour, which have been described as rebound effects

(Midden, Kaiser, McCalley, 2007). Also, resistance to new systems and negative experiences, for example

due to faulty automation or lacking user friendliness, has frustrated the high hopes of innovative technologies.

On the other hand, the effects of purely behavioral approaches have been very successful neither or mixed at

best (e.g. Weenig & Midden, 1997). One of the main reasons for the lack of success is that most

communication programs targeted the intentions of users, but largely ignored the technical context in which

consumption choices actually occur. Basically, we view energy efficiency and conservation as the outcomes

of multiple interactions between technological systems and human users. It follows that interventions that aim

to influence consumption behaviour should be concentrating on guiding interactions between users and

systems.

The linkage between technology and sustainable user behavior can be described by distinguishing four roles

of technology: (1) as an intermediary, where the technology used for attaining a goal defines the ecological

impact, although often surrounded by uncertainty; (2) as an amplifier, where technology amplifies the human

potential to attain goals, but at the same time it amplifies the use of resources (3) as a determinant, where

behavior is shaped and activated on the basis of the affordances, constraints and cues provided by the

technological environment and (4) as a promoter, where technology is designed to influence behavioral

choices (Midden, Kaiser & McCalley, 2007).

Although it would be worthwhile to consider each of these roles as a perspective to design persuasive

interventions that enhance sustainable consumption, we will focus in this contribution on the fourth role of

technology, that is the role of promoter.

2. Technology as promoter of sustainable behavior

How can people be motivated to use scarce natural resources in a sustainable way? In the search for effective

interventions we ask how persuasive technology can help to overcome some traditional limitations and make

motivational strategies more powerful. We discuss this role of technology regarding two foremost challenges

that policy-makers and psychologists face in combating the major environmental risks of CO2-emissions and

climate change. First the use of media technologies is explored, to see how they can help enhance problem

awareness. Second, we focus on interventions to change behavior and the ways technology can be used to

make interventions more effective.

2.1. Using novel media to raise risk awareness

Since the 1970s, worldwide numerous mass-media campaigns have been used to raise concern for the threats

to natural eco-systems and the urgency of action. Results have often been disappointing. Among the many

issues that have been identified, attention and processing issues form an important part.

Looking at attention rates, many mass-media appear not to be used by the general public. Traditional visual

media such as television ads and video-clips have also been used to stimulate environmental awareness and

conservation behavior. However, visual media are not more motivating per se for enhancing sustainable

behavior, in spite of their easy access and less demanding processing (Weenig & Midden, 1997).

More advanced multimedia technologies may add persuasive impact to the traditional communication of

transferring symbolic information (like text or speech) by inducing direct sensory experiences like sounds,

images, scent and touch that create ’presence’, the feeling of ’being there' in a mediated environment (see for

an overview IJsselsteijn, 2004). It may allow people to better conceptualize cause-effect relationships, such as

how an urban area would look and feel like without car traffic or how the world would be after serious climate

change. More recently, significant research efforts have been directed toward investigating the relation

between ’presence’ and emotional impact (measured through, e.g., galvanic skin response or heart rate

variability), where findings are supportive of the existence of such a relation, in particular in relation to fear-

inducing media environments (Meehan, Razzaque, Whitton, & Brooks, 2003). Research in the domain of

environmental risks has convincingly demonstrated the role of affect and emotion in risk perception (e.g.,

Slovic, Finucane, Peters, & McGregor, 2004,). Some experimental evidence is available which shows that

video images with emotionally charged content stimulate attention for climate risks and coping options. The

use of intrusive images and dramatic sounds to alert people were found to enhance relevant information

processing for coping with these risks (Meijnders, Midden & Wilke, 2001).

These studies suggest a new area of inquiry in which virtual environments can be used to offer new

opportunities for technology assessment by giving people pre-experiences of future technology effects or

newly planned environments and facilities, which will go beyond verbal descriptions or abstract

representations. Ongoing work in the Netherlands focuses on the cognitive and motivational effects on coping

behavior as a result of user experiences in a virtual polder environment, which is threatened by dike collapse

(Zaalberg & Midden, in preparation).

In sum, traditional media have had limited success in promoting environmental problem awareness, but new

multimedia technologies show more promise in this endeavor by offering new opportunities for creating and

enriching sensory experiences as a route to raising awareness of future and/or distant issues, to explore cause-

effect relationships and to experience environments that are not directly observable. However, despite the

possibilities offered by multi-media technology, raising awareness will not be enough to fight climate risks

and diminish the use of natural resources. In the next section we turn to the role of technology in

accomplishing behavioral change.

2.2. Using Persuasive Technology to Promote Energy Conservation Behavior

Prior to the 1990’s, experiments using electronic devices indicated that they might contribute to the efficiency

and effectiveness of behavioral interventions, but technology just wasn’t yet smart enough in most cases to

make these devices very successful. Psychologists have as yet merely touched upon the opportunities offered

by intelligent systems to promote (energy) conservation behavior. Most early work was done on the effects of

feedback on energy consumption in the home. Studies often used simple procedures like written messages

based on daily or weekly meter readings, while some researchers used electronic displays. Electronic modes

of feedback have been proposed to solve a number of issues related to written modes. First, electronic means

could provide feedback more quickly and frequently than written feedback, even continuously, thus making

the consequences of specific behaviors better available for the consumer. Second, electronic feedback could

be given at more central locations like the living room or the kitchen. Third, electronic feedback allows for the

use of multiple standards (e.g., personal and social), reference points (e.g., financial costs per hour, the

previous day or the upcoming month) and units (e.g., $ or emitted CO2). Fourth, written feedback provided

with a high frequency has been quite effortful and costly. Automation could make the feedback process more

efficient. Fifth, instead of the usual aggregated feedback at the household level, electronic feedback could be

source-specific (e.g., the airco or the cooker), evidently creating a closer link between feedback and action

(e.g. Wood & Newborough, 2003).

In sum, electronic means have made it easy to provide highly frequent feedback, which is more effective.

Electronic devices have also facilitated feedback on specific appliances, which appeared to be more effective

than general feedback. Goal-setting, added to electronic feedback, enhanced energy savings.

Almost all interventions were designed to communicate with subjects in a one-way direction. Modern

intelligent systems enable two-way interaction between user and system, which allows for more precise

targeting of tasks and for personalization. To illustrate, interactive systems allow for the implementation of

more refined goal-setting procedures and the provision of more specific information, not only to specific

appliances but to specific tasks as well. Interactive devices are still rare in the domain of (energy)

conservation behavior. Some studies, however, illustrate the potential. The present authors observed in two

earlier studies (e.g. McCalley, 2006) energy conservation results up to 20% using washing machines with a

user interface that allowed for interactive goal-setting and outcome feedback. During a series of twenty

washing tasks, users received immediate feedback each time they made a choice for a washing program to

carry out a particular task. Subjects with either self-set or assigned goals saved more energy than subjects

without an explicit goal.

Applying intelligent agent technology that learns from the users and interactively communicates on a personal

basis could enhance the power of supportive systems. We use the term agent to refer to a piece of software

that can be considered as an autonomous creature able to perform tasks with more or less intelligence and

autonomy. It can be made visible in many ways through virtual or physical forms of embodiment (e.g.

Diesbach & Midgley, 2007). An agent system could be able to frame outcomes based on the current context or

user, or to encourage the user to make certain goals more explicit, and even make suggestions on how to act or

guide a user to a decision. In this role, intelligent agents may become persuasive social actors, rather than

simple tools (e.g. Fogg, 2003). In a very recent study (Midden & Ham, 2008) demonstrated that social

feedback from a physically embodied agent, an iCat (Philips company), resulted in significantly more energy

conservation behavior on the same washing tasks than the factual feedback like provided in McCalley and

Midden study.

3. Conclusion

In this paper we discussed behavioral interventions on enhancing sustainability with a focus on human-

technology interactions. Four roles of technology have been suggested and the role of promoter of sustainable

behavior has been explored in greater detail. Our review reveals that persuasive technology has much to

contribute to the design of effective motivational interventions. It helps to raise awareness of future or distant

issues, such as the vast melting of polar ice, or to lower thresholds for change, for example by making it

possible to experience a building not yet constructed or to explore cause-and-effect relationships such as the

effects of ventilation on air circulation in the home. Technological assistance may go beyond the level of

specific appliances or systems and direct energy use. For example, the application of computer and robot

technologies to domestic appliances) will be able to monitor multiple sources of energy use and support home

energy management. Such systems will offer advice on saving options taking account of personal lifestyles

and will even be able to support strategic decisions like investments in equipment and home renovation.

Sustainability requires joint efforts in various social groupings (e.g. household, neighborhood). Persuasive

systems may also be able to touch this social dimension of sustainable behavior for example by coordinating

contributions.

Interventions that aspire to integrate psychological with technological means form a challenging perspective.

We believe, however, this effort to be most worthwhile on the route to a society that makes sustainable use of

its natural resources.

4. References

Diesbach, P.L. & Midgley, D.F. (2007) Embodied Agents on a website: Modelling an Attitudinal Route of Influence. In: deKort, Y.

Ijsselsteijn, W., Midden, C.J.H., Eggen, B, & Fogg, B.J. Persuasive Technology,

Ehrlich, P. R., & Ehrlich, A. H. (1991). Healing the planet. New York: Addison-Wesley.

Fogg, B. J. (2003). Persuasive technology. Amsterdam: Morgan Kaufmann Publishers.

Meehan, M., Razzaque, S., Whitton, M.C., & Brooks, F.P. (2003). Effect of latency on presence in stressful virtual environments.

Proceedings of IEEE Virtual Reality 2003, Los Angeles, California, 141-148.

McCalley, L. T. (2006). From motivation and cognition theories to everyday applications and back again: The case of product-

integrated information and feedback, Energy Policy, 34, 129-137.

Meijnders, A. L., Midden, C. J. H., & Wilke, H. A. M. (2001a). Communications about environmental risks and risk reducing

behavior: The impact of fear on information processing. JASP, 31, 754-777.

Midden, C.J.H., Kaiser, F.G., and L McCalley, L.T. (2007) Technology’s Four Roles in Understanding Individuals’ Conservation of

Natural Resources. Journal of Social Issues, Vol. 63, No. 1, 2007, pp. 155—174.

Midden, C.J.H. & Ham, J.R.C.(2008) The persuasive effects of positive and negative social feedback from an embodied agent on

energy conservation behavior. Unpublished manuscript. Eindhoven |University of Technology.

Slovic, P., Finucane, M.L., Peters, E., & MacGregor, D.G. (2004). Risk as analysis and risk as feelings: some thoughts about affect,

reason, risk and rationality. Risk Analysis, 24, 311-322.

Weenig, M. H., & Midden, C. J. H. (1997). Mass media information campaign and knowledge gap effects. Journal of Applied Social

Psychology, 27, 11, 945 – 958.

Wood, G., & Newborough, M. (2003). Dynamic energy-consumption indicators for domestic appliances: Environment, behaviour

and design. Energy and Buildings, 35, 821-841.

IJsselsteijn, W. (2004). Presence. Doctoral dissertation, Eindhoven University of Technology, Department of Human-Technology

Interaction, the Netherlands.

Zaalberg, R. & Midden, C.J.H. Adaptation and prevention. Real and Virtual Experiences with river flooding in the Netherlands. In

preparation.

About the authors: Cees Midden is professor of Human Technology Interaction and chair of the Human-Technology Interaction group

Pervasive persuasive technology and environmental sustainability (Workshop)

Abstract and Figures