Citizen jury about trust in data systems and models
Context and background
Trust is crucial for dealing with ‘an uncertain and uncontrollable future’ (Sztompka 1999: 25) or situations involving risk. The ways in which organisations handle personal data are arguably uncertain and uncontrollable, and the broader data ecosystem in which they are situated is characterised by insecurity. In a survey in the UK, doteveryone found high levels of concerns about the risks associated with organisational uses of personal data – between 91 and 95% of people thought it was important to know how their data was being used, to have control over how much data they share, and to know their data is secure (doteveryone 2018). And yet, trust in this domain is remarkably low. A 2014 Royal Statistical Society poll found there is a ‘data trust deficit’ – only 6% of citizens reported high levels of trust in Internet companies and overall, trust in institutions to use data appropriately is lower than trust in them in general (RSS 2014: 1).
In some research into trust in data practices, it is assumed that there is a link between trust and understanding. The doteveryone research cited above concludes that without understanding, people are not able to make informed choices about technologies and ‘it is likely that distrust of technologies may grow’. In contrast, in qualitative research into how people manage data in their daily lives, Pink et al (2018) show that trust has an affective dimension and its relationship with feeling is therefore as important as its relationship with understanding. Research in Human Data Interaction identifies legibility as a key enabler for agency and as a pillar for building trusted relationships with data driven technologies, but the relationship between understanding and feelings in producing trust remains largely unexplored. We explored the complex interaction of knowledge, feelings, and trust as they relate to current and future data practices by holding a citizen-jury-style workshop.
Trust is elusive in our current data-driven economy, and yet it is necessary for data-based systems to function in a sustainable way and with the support of the citizens on whose data it relies. In this project we will explore public understanding of future data-driven systems, perceptions of AI futures and criteria for trusted data systems using qualitative methods, in the form of focus group research.
This research extended the survey we conducted. By deliberating with participants on future data management scenarios, we were able to access their thoughts and feelings that cannot be accessed by a survey. Compraing methods in this way, we were also interested in exploring the best methods for engaging publics in conversation about these matters, to ensure that future data driven system fit with the ethics and values acceptable to wider society.
What we did: citizen jury workshop
We ran a citizen-jury-style workshop with diverse citizens to explore their trust in data-driven systems and data management models. Citizen juries are a policy making tool where diverse citizens are brought together to debate a complex issue of social importance and make a policy recommendation. Citizen juries seek to facilitate participatory and deliberative democracy, in which citizens work together not simply to provide opinions on important issues, but to synthesize those opinions together and present answers to complex questions. In order to facilitate an informed discussion, experts are brought in to present evidence and a facilitator is used to ensure that issues are approached from multiple angles. In citizen juries, citizens are thus given a chance to contribute their informed opinions about issues that could materially impact their lives.
Citizen juries often extend over several days, but with limited resources available, we had to keep our workshop to one day. We incorporated key citizen jury elements of deliberation and informed decision making into our workshop. We included presentations from domain experts in data-driven systems and data management models. Expert 1 introduced the benefits and harms of data-driven systems that have been identified by other experts. Expert 2 introduced five data management models, including data trusts and personal data stores, and their advantages and disadvantages. 12 participants, recruited by XXX, engaged in a facilitated discussion where they considered the benefits and weaknesses of different approaches. Participants decided as a group the most important criteria for the design of trustworthy data-driven systems and data management models.
What we found
Findings from the citizen jury workshop will be available soon.