Going hand in hand: data and design
Written by Cat Drew & Laura Williams, with Jon Flint, Dr Makayla Lewis & Phillippa Rose
2020 will be remembered for one major thing, but one of the side effects of the pandemic was supercharging our reliance on digital, and the data and machines that drive it. And continuing a long-standing narrative around data & AI, we have seen huge successes during the year: the Taiwanese test and trace systems (that worked), data visualisations which evidenced the impact of traffic reduction on air pollution, a huge uptake in home delivery with prioritisation given to key workers and those most vulnerable. But — where badly designed, there are also concerns. Test & trace systems that haven’t worked, or where data has been kept and misused, the greater uptake in AI for health-related decisions about who to triage and treat — although not covid-related, the A-level exam result fiasco, which has done much to build mistrust among a generation of young people, before — for some — a first term at university like no other, and for others, an incredibly challenging job market.
Digital and data run through everything we do, and ethics is increasingly becoming a core part of the design process. Design and ethics feel symbiotic. At their core, they are both about understanding how things should work in a way that people want and are comfortable with. Good design needs technology to make things work better, faster, smoother. And tech needs good design to make sure that things are adopted, inclusive and make lives better, for all. Sacha Costanza-Chock’s book on Design Justice is as brilliant and comprehensive as you’ll get on this.
At the beginning of the year (which honestly feels like a lifetime), we ran an event looking at the potential for AI within the planning, health & retail sectors, and the need for design to make sure these technologies were inclusive. Since then, we’ve run a masterclass on AI & design at Service Design Days in Barcelona, designed the updated Government data ethics framework, worked with Social Tech Trust to co-design a scaled-up entrepreneur programme, hosted an event on best practice digital community engagement and are working with the Centre for Digital Built Britain to develop a digital learning curriculum for young planning professionals. And that is on top of all the work that our Design Associates are doing outside Design Council.
As part of our Design Associate Community of Practice, we brought them together to share insights from our collective work, which we also share openly here.
We started with the updated Government’s Data Ethics Framework which we had visually designed, and explored how design was a core part of making data projects ethical, as well as wider opportunities for design:
Before the project: public awareness and understanding of the potential of digital design, data and AI
We know from the research conducted by Ipsos-MORI which informed the first iteration of the Government’s Data Ethics Framework in 2016, that before people can start to weigh up the detail of data ethics, they first have to a) understand that data science is possible, and b) that the goal of the project is worth the level of personal intrusion.
Clear graphic communication can help on both of these fronts. The Royal Society has a clear explanation of what AI and machine learning are, or this a great visualisation of how it works in practice (showing how a machine can predict whether houses are in San Francisco or New York). During a trial for how GPS could indicate less crowded parts of the tube, Transport for London created posters which mocked up how you might experience it on an app, so you could immediately see the benefits.
More provocatively, design also has an important role to play in raising awareness of the extent to which our lives are controlled or influenced by data, so that we can start to question it. Jon Flint (one of our Design Associates) is part of the Cached Collective — an international group of creatives that use their expertise in design, prototyping, narration, visualisation and embodiment to create outputs that raise questions and awareness about data and privacy in a tangible, friendly and immersive way. Their projects have created objects (e.g. mirrors or receipts which reflect back your digital self as perceived by social media algorithms) which are a user-friendly wake-up call that invites you to think critically about your online behaviour. There is a whole collection of similar projects here.
During the project: human-centred design and prototyping
Good data ethics means starting with a clear public need or goal, and a central part of a design process is to spend time with people to understand what that is, get to the root cause of the issue, and then design the technology or data in support of that. In addition, Anne Dhir from Snook also talks about designing for the positive (e.g. ensuring people receive benefits/free school meals that they are entitled to) rather than the negative (e.g. collecting fraudulent tax evaders).
Even if the data model is constructed fairly, the ethics is around the policy decision that is taken as a result of that. It is really great that the new data ethics framework has a new final element which is “Evaluate and consider wider policy implications”, which asks the policymakers to check that the insights are being used for the original intent. Prototyping can help here, as it can start to make transparent different potential uses and unintended consequences. This can happen even in a very simple way. Phillippa Rose and Dr Makayla Lewis (two of our Design Associates) and I ran an AI & Design masterclass at this year’s Service Design Days conference that asked participants to first map out a user experience journey (for example visiting a loved one in a care home), and using a set of AI method cards to imagine how these could be positively or negatively used to change the experience. You can access the miro board here.
As a group, we are clear that AI and machine learning should not replace professionals, but allow them to do their work better, and the more positive scenarios from our workshop envisaged healthcare professionals or planning professionals being able to spend more time with communities, building trust and relationships. We need to continue designing the human, relational aspects of a service and a place, with technology that supports that, not the other way round.
More widely — enabling and embedding ethics in education, the design sector and other organisations
As with design, ethics is not just a toolkit that you can use on a project and then wash your hands. And nor should it be the preserve of a specialist or a team. It needs to be embedded within organisations. Makayla’s work has focused on the uptake of AI within organisations, and she says that the following areas can impact it.
● Culture: an environment that is resistant to change, or the inability of some professionals to see that the disruption caused by AI can be beneficial.
● Fear: linked to the above, for some, the large unknown surrounding the role AI will play in the future, for example whether it will replace jobs, or not being able to predict where its use might spread to.
● In-house talent: shortage of professionals with the knowledge and experience required to design and implement the required infrastructure and organisational change
● Strategic approach: limited planning and understanding of AI at the strategic level.
Both our workshop and masterclass ended in a discussion about what this means for design.
● A greater understanding, simplification or de-mystification of what AI is, and potential positive uses and negative consequences, applies to designers (see point 1) applies to designers as well as the general public.
● We need more ‘specialist generalists’, whether those are designers who understand tech and ethics, or technologists who understand design. Either way, these are people who are translators and therefore can bring together different perspectives and multidisciplinary teams, and support organisational culture change as set out above.
● Design education needs to include more around digital, tech and ethics, and about the importance of being able to translate and work with different designers, data specialists, ethicists, technologists, professionals and of course, citizens.
The Government’s Data Ethics Framework’s primary audience is policymakers, and we discussed how we could take the core principles and adapt and visualise them for a design audience, as a common design standard. Or indeed, share some of the other (and increasing) AI & design tools and resources that are out there, which we’ve included below.
To wrap up
Design is key to making data projects ethical as it:
- Helps make the invisible, visible, through visualising and mapping activities, explaining what AI & machine learning are, the power of what they can do, and the extent to which they shape our lives
- Enables a better understanding the current and future potential impact of intended and unintended consequences for each design decision, through a considered, conscious and reflective practice
- Frames design considerations in the context of bias and aiming to correct these — or better — design them out, building collective empathy into the process
- Encourages collaborative working with a mult-disciplinary team of specialists and experts, to provide advice, ideas, guidance, governance and accountability — building capability and learning through the project and into future projects
Data and design go hand in hand. Design needs to be a central part of any data project, and data & ethics should be a central part of any design practice.
Other resources:
Phillippa’ reading list of all things data, design & ethics
Cat’s TedX talk on data x design
Jennifer Panting’s blog on Design Council’s AI & Place workshop
