The Data Justice Lab held an international multi-stakeholder workshop at Cardiff University’s School of Journalism, Media and Culture (JOMEC) on 7th June 2019, as part of its ongoing research on participation in a datafied society. It brought together more than 30 participants from organisations across government, civil society and academia, mostly from the UK but including the US, Canada, Spain and Iceland.
The purpose of the workshop was to explore different perspectives on civic participation in data systems and the challenges within different decision-making and governance models. The workshop was divided into three sessions: Involving the Citizen, Institutional Practices and From Infrastructure to Advocacy: Citizens Interventions.
Session 1: Involving the Citizen
In the first session we explored different methods of including citizens in both data and political systems. We heard from speakers with experience of setting up citizen juries as well as representatives from the New York algorithmic task force and the Icelandic crowd-sourced constitution process. Discussion largely focused on the challenges in deliberative processes such as citizen juries and the barriers to meaningful participation. There was some agreement that to be meaningful these methods must enable citizens to have tangible input into political decision-making with funding set aside for policies. Questions were raised, though, regarding the legitimacy of mini publics.
Session 2: Institutional Practices
The second session investigated governmental data strategies and efforts to advance participation. Attendees from government pointed to the diversity of strategies, with some more advanced than others. Some authorities are fearful of how their projects will be received and others are more actively engaging with the public. The roles of new policy frameworks and institutions were highlighted – incl. the General Data Protection Regulation (GDPR), the Centre for Data Ethics and Innovation, and Information Commissioner’s Office (ICO). The topic of representation (particularly of marginalised communities) in digital systems was prevalent, with additional questions raised around public consent.
Session 3: From Infrastructure to Advocacy: Citizens Interventions
In the third session we explored broader strategies and practises to enhance democracy in an increasingly datafied society. Presentations addresses of ‘citizen audits’, subject access requests, policy advocacy and technological design, among others, across a range of international contexts. The discussion addressed the political contexts which made certain actions possible or not, ranging from the presence of street movements to having a sympathetic mayor. A discussion on data ownership highlighted different conceptions of what this means in practise and its implications, with concerns raised about the marketisation of data. Rights frameworks and stewardship were suggested as alternative models.
The debate uncovered different approaches to participation, from a limited and consultative involvement of citizens in some decisions to a more thorough rethinking of technology and administration in order to incorporate participation into their very design, to questioning the use of technology in the first place to avoid risks of further marginalisation of already disadvantaged communities.
Workshop themes
- Defining participation
There was no clear consensus as to what meaningful citizen participation encompasses, with workshop participants pointing to a range of different challenges and offering different conditions. Some participants stressed the importance of tangible outcomes. Others took a more normative approach by emphasizing that participation should be inclusive and representative, which was deemed a potential challenge for deliberative methods like citizen juries. There was disagreement here as to how representation can be best achieved; for example, whether all voices need to be accounted for, or whether more weight should be given to social groups most impacted by datafication.
With respect to the implementation of data systems, several speakers highlighted that citizens need to be participating at an early stage in the process, ideally before a system has been developed. It also became clear that deliberative methods are just one approach in an ecosystem of participatory methods, and that there is a danger in seeing participation as an advocacy exercise or a means of obtaining the impression of public buy-in to an issue.
- Data rights
Some speakers expressed concern that data rights are not enough to mitigate or address data harms, suggesting, for example, that “rights based discussions are too single minded, too individualized” and that there is a need to go beyond privacy towards collectivizing knowledge. This contrasted with those who saw potential in data protection, such as the accountability potential in data protection impact assessments as well as the value of subject access requests as a tool for enforcing compliance on organizations and auditing their data practices. Some suggested that data rights can be used to encourage participation in data systems, such as tools like My Data Done Right which, if used collectively can be powerful and force GDPR compliance on organisations. It was suggested that data rights do connect with civic participation, for example by using subject access requests to perform citizen data science experiments. It was claimed there was a need for non-personal data regulation as the GDPR’s remit may fall short in some cases of harm if it cannot be tied to personal data.
- Inequality and representation
Throughout the three sessions concerns were raised that data systems are exacerbating existing inequalities and that the impact on different groups is uneven. However, as these observations were linked to structural and systemic inequality there was also a recognition from several participants that doing away with the technology itself would not alleviate harm or struggle. Some participants spoke of the challenge of how best to include weaker voices in mini publics (deliberative exercises). According to one speaker, deliberative methods are limited by the fact that they “are missing the broader networks of civil society” such as activists, environmental groups and mothers’ groups. Others with experience of mini publics acknowledged that recruiting citizens to juries and panels is done on a cross section basis rather than engaging impacted communities like immigrants or the homeless. This was tied up with the view that deliberative approaches are “equity based spaces” where everyone’s voice is important.
Further, concerns were raised about civil society not adequately addressing this inequality or representing those most affected. For instance, one academic noted that digital rights groups are limited because “they do not understand what is happening with more marginalised communities”, leading such communities to be skeptical of issues that are predominantly framed by data protection and digital rights. Similarly, a digital rights litigation spokesperson said that more grassroots organisations need to be brought onto litigation cases in order to address affected people. A civil society representative said there is a “need to talk to the groups that actually represent these particular interests so they can understand the nuance of the issue.”
- Collective models bring public benefit
For several participants collective models offered a better approach to governing data because they were seen to provide wider public benefit while also having the potential to counter power imbalances between corporations and citizens. Some participants talked about data as a social good and a “common pool resource” that could be controlled by and shared within a community in order to solve social problems, and which could be used to build alternatives to the current platform economy model of surveillance capitalism. Other participants thought that a collective model could correct power imbalances in the platform economy by treating data as a community resource that can be exchanged or “traded” with organizations in return for access to services. However this was met with the criticism that data then becomes valorized in market terms.
- Framing & impact
New democratic processes may not necessarily have a tangible positive impact and may need integration with existing decision making processes. Some participants expressed concerns that some deliberative tools could be used as a box ticking exercise to give the impression the public is being engaged with little or no change to existing arrangements or trajectories. One participant claimed participatory budgeting is a particularly interesting tool as it has a definite material output built into the process itself, potentially side-stepping some of these issues.
Tied to this question was that of framing. It was claimed that an unclear, poorly defined, or even cynically constructed process could be rendered ineffectual. One participant claimed that if a process lacks a clearly defined output then that is a red flag for it being a flawed or misused process. Another claimed, “if you ask the public a stupid question they’ll give a stupid answer”. A particpant with experience of constitutional reform said in their experience established power will often try and undermine a process to maintain its position. Money was also mentioned as a limiting factor, not only in the funding of a process but also in that it can decide which organisation has the ability to comission a process and, therefore, frame it, set its boundaries, etc. It was claimed that a well designed process could avoid many of these shortcomings but that, nonetheless, we should not expect any one process to produce perfect results and should always seek to refine our methods and models.
- Reform vs change
Many participants accepted that while many of the processes being discussed (i.e., mini-publics) were far from perfect they were nonetheless an improvement on our current options for democratic engagement and so should be championed and expanded if possible. Others (without wholly dismissing this position) claimed that our issues were much more fundamental and required a wider restructuring of society. This was reflected in different approaches that participants took, with some focusing on what government and other institutions can do, whilst others expressed a desire for more bottom-up responses. It was claimed that too often solutions and policy decisions are framed as a technical fix, whereas the issues that need fixing are much deeper and related to how we run our societies more generally. Those expressing sympathies to the top-down side of this divide did not outwardly disagree but stressed that they thought it necessary to achieve concessions and advances in the here-and-now, such as through some of the more modest projects that had been discussed throughout the day. It was claimed that all levels of community organising should be engaging with issues surrounding data.
- Public knowledge and willingness to engage
A recurring issue raised was the difficulty of engaging with the public. A government attendee framed it as a lack of willingness on the part of the public to engage. It was pointed out that it is difficult to get media coverage on an engagement exercise, perpetuating a lack of public knowledge, a catalyst for engagement. Participants noted that civil society and community groups need to be involved more directly. It was suggested that people’s lack of knowledge about an opportunity to engage with institutions may not be the only bottleneck, but people might not be aware of the importance of the issue being engaged on. Technological issues can seem very distant and alienating, it was claimed, and this is a barrier to engagement. A novel suggestion was to invert the dynamic, such as running a “reverse citizen’s jury” where lay members of the public give their opinion as citizens to experts in fields such as data and artificial intelligence. The importance of nuanced and detailed discussion was raised (with Brexit raised as the model of public discourse to avoid).