Design for the change before behaviour

A blog series on how AI agents could be designed to enable positive social and environmental change

Gyorgyi Galik
6 min readOct 21, 2020
Illustrations © Niklas Hagemann (niklashagemann.com)

Currently designed primarily for home automation, home entertainment, and shopping, there are 52 million+ Google Nest devices and 100 million+ Amazon Alexa devices that have been sold in people’s homes, offering services like ordering a pizza, calling a taxi, or playing people’s favourite music. Google’s voice-driven AI helper, Google Assistant was expected to be on a billion devices by 2019.

Piggybacking on this ‘trojan-horse-like’ moment deployed in our most personal spaces, I was interested in designing a more socially and environmentally minded adaptation to current AI home assistants.

My practice-led PhD in Innovation Design Engineering at the Royal College of Art was exploring how different technology enablers could be designed and applied to sustain public engagement around the climate crisis and the reduction of energy use. It also offered novel approaches to current demand-side response strategies — asking people to reduce or shift their energy consumption when electricity demand from the grid exceeds supply.

The PhD set out to design technologies that increase people’s agency in taking meaningful social and environmental action, individually as well as a collective. The aim of this was to aggregate small, individual action and challenge the perception that small changes by individuals do not matter and are somewhat futile when it comes to the inaction of a whole world of other people.

This blog series shares the learnings from the PhD and aims to offer evidence-based practical alternatives to current approaches and narratives of smart homes, automation and ambient intelligence and suggests how design research practices could expand this currently narrow design space and contribute to low-carbon and low-pollution innovation.

The work has value for those ' designing for agency’ and working in design for behaviour change, as well as for designers and other professionals involved in energy research, developing virtual assistant AI technologies and improving the interaction with AI agents.

After discussing the necessary political and systems changes that need to be put in place first, a number of social psychology and behaviour change insights are applied in practice to understand the most appropriate ways of helping people transition to low-pollution lifestyles.

Moving beyond the mitigative aspects of current technological interventions — which won’t be enough to address the complexity and severity of these challenges alone — a new set of technology design principles will be shared and detailed for other designers and developers to use (discussed in an upcoming blog post on Designing technologies that could achieve a greater collective impact).

The PhD set out to shift the focus from pollution observation and monitoring (sensing, mapping, sonifying and visualising pollution that has already been produced) to pollution prevention (preempting pollution before it is produced in the first place).

It explored the following research questions:

(1) Could a connected technology be designed to engender preventative behaviours and afford a more proactive role for citizens in making and/or supporting the decisions that prevent pollution in cities? (bottom-up, individual change)

(2) If networked, could a novel interaction be designed to stimulate societal demand for environmental regulation (top-down, systems change) and to aggregate the small impact of individuals to achieve a greater collective impact?

While keeping in mind that current AI home assistant devices still have considerable technological limitations, in the first design experiment, in collaboration with Becky Jones, a new socially and environmentally minded skill was developed and tested for an existing artificial intelligence (AI) assistant technology.

In the second experiment, following in Delfina Fantini van Ditmar’s footsteps and applying the Wizard of Oz interaction technique, I took on the role of an AI agent, to better understand my participants’ views on sustainability, energy use and what they value and prioritise in their daily lives, and what activities they might be willing to change.

I was deeply inspired by Delfina’s 2016 PhD thesis, “IdIOT : second-order cybernetics in the ‘smart’ home” in which she takes on the role of “a smart fridge software and collects both quantitative and qualitative data” [1] to question the algorithmic logic behind Internet of Things (IoT) technologies. She looks at [2]:

how human lives are represented within the quantified approaches inherent in current notions of ‘smart’ technology, derived from Artificial Intelligence (AI), and characterise this as the Algorithmic Paradigm.

She sets out to understand the logic behind algorithmic conversations and explores “questions of how complex, lived, human experience is oversimplified in the IoT” [3]. Her work emphasises that “smartness is relational” and calls for “a shift in perspective to create more meaningful interactions with devices in the smart home” [4].

Climate Pal prototype (left) & Wireless Sensor Tags kit (right) monitoring humidity-levels in the shower

In the third experiment, a new design method and custom-built, digital assistant or social companion known as Climate Pal (CP) was developed in collaboration with creative technologist, Tim Brooke and with a group of participants, demonstrating the potential for AI agents and home assistant technologies to help people change behaviour, increase the degree of participation in reducing pollution in cities and help people articulate their own agency in complex environmental matters.

My participants in the final experiment listening, sharing stories and taking collective action

To enable collective environmental action, the fourth and final experiment offered a new alternative to further increasing participants’ individual agency by connecting them to a group of like-minded people through a network of AI home assistant devices.

This new approach tested whether the social aspect of the intervention made participants’ engagement and the changes in their behaviour more durable than if the intervention were to have stayed solely focused on individual action.

Instead of designing for the behaviour to change, the experiments focused on designing for the change before a behaviour was even performed. By breaking participants’ daily routine with the help of a new technology enabler, well-designed and deliberately staged disruptions or triggers helped participants take a pause and think, enabling shifts in their current behaviour and in the uptake of more sustainable behaviours.

The design experiments characterise a more prevention-focused, design research-led, and agency-sensitive approach to pollution reduction and to behaviour change.

While current technologies often design with a focus on the individual, the outcomes of the design experiments point to the need for new technologies that are deliberately designed to involve more people to simultaneously interact with a technology enabler and work towards a shared goal. The evidence emerged through the PhD suggests that technologies that account for and support people’s core values and desire for social connection and belongingness, for being challenged by, and learning from, others; for doing good for other people; working towards a shared goal; and becoming part of something greater than themselves will likely be more successful in enabling sustained engagement and more meaningful change at scale.

A set of new technology design principles that are introduced in the blog post Designing technologies that could achieve a greater collective impact aims to offer an alternative to the current solely observing and visualising approach of e.g. smart meters and energy use, air quality sensing technologies and clean air route finder apps, and to enhance people’s agency in reducing pollution in cities.

25 design principles that will be detailed in the blog post ‘Designing technologies that could achieve a greater collective impact’

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

References:

[1] Fantini van Ditmar, D. (2016). IdIOT : second-order cybernetics in the ‘smart’ home. pp.124–125 Available at https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.690638
[2] (ibid., p.1)
[3] (ibid., p.1)
[4] (ibid., p.1)

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — —-

Gyorgyi Galik is a London-based innovation designer, design strategist, and environmental advocate. She is a Lead Programme Manager at Design Council’s Cities Programme. She has recently passed her PhD in Innovation Design Engineering, School of Design at the Royal College of Art. With a background in social design, behavioural science, and environmental health, she has more than a decade of experience delivering experimental research, design and technology projects in the corporate, governmental and non-governmental sectors. She has designed, led and managed a variety of projects in North America, Asia and Europe — working in organisations ranging from small startups and individual communities to large corporations and city governments.

--

--

Gyorgyi Galik
Gyorgyi Galik

Written by Gyorgyi Galik

Innovation designer, design strategist, and environmental advocate — interested in collective action, cities & the climate crisis https://twitter.com/GyGalik

No responses yet