Since the Centre for Data Ethics and Innovation (CDEI) was established two years ago, engaging with the public has been a core component of our work.
For the CDEI to advise on best practice in responsible innovation, we need a deep understanding of the way in which the UK public want data-driven technologies to be used now and in the future.
There are challenges to public engagement in data and AI. The subject matter is technical, and policymaking in this area can involve particularly complex trade-offs. But we believe it is vitally important, and it is something we committed to in our original Terms of Reference and 2 Year Strategy. The recently published National Data Strategy reiterates this commitment.
To inform our review into the governance of online targeting, we conducted deliberative public engagement. We engaged over 140 members of the public over a two-day period to inform them about how targeting systems work, understand how they want online platforms to be governed, and work through the trade-offs involved in new models of governance.
For the CDEI’s recently published review into bias in algorithmic decision-making, we wanted to understand public attitudes towards the risks of bias and discrimination in algorithms. Here we used a different public engagement methodology, in addition to survey research, running an experiment with the Behavioural Insights Team to uncover attitudes towards the use of data and algorithms in financial services. People took part in an online experiment to explore how behaviour changed according to how (imaginary) banks used data to decide how much to lend to people, from which we could infer the public's views on algorithmic decision-making. We are now using a similar methodology to test people’s perceptions of control over the use of their personal data, and explore how this could be increased.
From the start of the COVID-19 outbreak, we have sought to understand public attitudes towards the use of technology to tackle the pandemic with regular surveys to understand attitudes and awareness around the use of data and digital technologies in the UK’s COVID-19 response. We repeated these surveys over a six month period, and will be publishing a paper in early 2021 summarising these findings, reflecting on trends across the population and what this means going forward.
We are now working to increase the amount of public engagement work we conduct and bring together all our public engagement work from across the CDEI. To do this, we are setting up a new Public Attitudes Insights Team. This will ensure the CDEI’s work is informed by a deep understanding of public attitudes towards data-driven technology, as well as engaging with policy experts in government, industry and civil society. These insights will be used by the CDEI and policymakers to monitor the overall landscape and inform our advice. Insights on specific use-cases will also help to inform our projects - so that we are able to advise on the practical steps required to ensure that data-driven technologies are being developed in a trustworthy way.
This new team will make use of a range of methodologies, allowing us to engage the public in the way most appropriate for the research question at hand. An understanding of each methodology’s merits and limitations will allow us to use these tools together to build a richer understanding of public attitudes. This will include capabilities for survey research, deliberative dialogues and focus groups, analysis of social media, and behavioural insights / gamification tools.
There is lots of great and innovative work being done, both inside and outside of government, to understand public attitudes towards the use of data and AI, and we are committed to working in a collaborative and transparent manner. That’s why earlier this year, we set up a new ‘Public Attitudes to Data and AI’ (PADAI) network for cross-Whitehall organisations involved in data policy. We want to make sure insight and best practice are shared across government so that the value and impact of public engagement work is maximised. We will also be regularly publishing data and analysis on our website.
If you are doing research into public attitudes towards data and AI, or are interested in collaborating with us, we would be pleased to hear from you. Please email firstname.lastname@example.org.