Data Governance Deep Dives #1: The cooperative model

By The Data Economy Lab Team

December 2nd, 2020

Image sourced by

About Data Governance Deep Dives

Data Governance Deep Dives, organised by Aapti Institute and Digital Commons Network, are a series of closed-door events that will bring together practitioners, researchers, academics, and other experts working on models to harness the value of data with the aim of redistributing power in the digital economy. Following the Chatham House rule, the purpose of these informal and intimate discussions is to explore the different decisions, structures, technological choices, and the lessons we as a community can draw from them. We hope these conversations can help move data stewardship out of the theoretical realm into reality, and examine its impact on communities and more broadly, the data economy.

Health Data Cooperative : Learnings so far, and challenges ahead

Beginning with insights from a cooperative that provides a platform for individual users to store their health data and control the use of this data – the experience of incorporating citizens into their health data management was shared. At the outset of this model’s  journey, it was found that while regulations upon data management may have been adopted, mechanisms of recourse and acquiring control over health data are not easily accessible to citizens. Further, most individual health data was scattered across sources – hospitals, insurance companies and smartphone apps. Yet with growing digitisation, digital tools available to citizens have increased – changing their capacity to participate in their health data management. This led to an initial study: how can citizens be motivated to participate in a collaborative, community-led model to govern and manage health data?

This feasibility study on health data management discussed was focussed on breast cancer, finding that innovations of processes were not as well as researched as other overarching studies – owing to a lack of proven monetary returns on the same. As such, the research agenda on the feasibility study was citizen driven – to understand the universal benefits perceived by citizens, terms of sharing they required, and entry barriers to both citizens and agents within the cooperative model. Conclusions from the research included some essential data governance principles for the cooperative:

  • Conditional Donation : Ensuring citizens the right to decide conditions under which they want to donate data.
  • Motivational Incentives : Citizen incentives are key to achieving significant data samples for research. This model guards against putting individual incentives ahead of the ‘common good’.
  • Collective Benefits : The use of citizen data should generate clear and unequivocal benefits to society.
  • Rights Management : Establish structural and technical mechanisms that both guarantee collective benefit, and manage citizens’ conditions, concerns and rights.

To better understand citizens’ conditions of donation, motivational incentives, and their possible propensity toward being a part of and helping govern a health data cooperative – a further study was conducted. This study offered various data sharing scenarios and allowed citizens to give their  opinion on diff models – would they donate or would they not? Interestingly, it was found that no incentive was changing opinions once held – but other key factors determining donation conditions were found. For example, citizens were most likely to donate their health data to aid chronic and rare illness research. Based on these findings, further principles were developed.

Questions that remain 

The idea behind cooperative citizen governance over data is partly trust – it is a structure that allows for community agency and control over their data, and the opportunity to spur research that is beneficial to the community. But ways to motivate people to participate actively and help realise and raise the strength of a cooperative is something that is still difficult to navigate. Perhaps it is the case that the more defined and narrow a cooperative’s purpose is, the more people are willing to share and engage.

One way to expedite this is to reach out to other cooperatives, publish material that speaks to community governance and ask questions about the current and possible governance of data. Communicating the value of data, and of aggregated data, toward the community’s interest is key to this process. Collective improvement of health is a strong incentive – the idea that data can be a force of transformation in this space may be an important motivation.

On the governance end of things, questions about individual roles in governance are challenging terrain. Structural considerations for how people are represented in a cooperative while accounting for the needs of other stakeholders crop up anew. Does the cooperative assembly then consist only of individuals or institutions as well? Should those who donate be considered automatic cooperative members or are they just donors? And further, how does the cooperative establish a fluid, non-conflicting public dialogue with research institutions about their data needs while preserving community interest?

The ‘value’ of data: What constitutes wrongful incentivisation, and what design choices can affect this in various contexts?

When thinking about data as a raw material to spur radical transformations to people’s live experience – it is important to consider how this value is communicated. The use and sharing of personal data is human rights-parallel, and the ‘wrong’ kind of incentives or value articulation may commodify individuals data – which would not be theoretically dissimilar from commodifying individuals themselves.

To address concerns of this nature, it is important to instill structural, technical, and governance mechanisms within a cooperative that are rights-preserving. Federated machine learning, for a technical example, is collaborative in its foundation and can help analyse data without taking it out of personal repositories. Synthetic databases and effective anonymisation serve similar protection.

Some of these questions are not mere challenges as much as they are design choices that require a level of ethical deliberation to decide upon – they are also incredibly context-specific. Another cooperative that attended the Deep Dive, for example, works to empower gig workers with free analytical insights that then also help inform local governance decisions. In the case of this cooperative, the board always has a majority of the cooperative – since it is locked into the bylaws and locked into member registrations, this control acts as the primary interest-preserving guardrail for members.

On the incentive front – this cooperative for gig workers also offered that within their model, gig workers receive dividends on the sale of their data insights to external agencies. In this scenario – when considering commodification of peoples’ data – it can be argued that massive companies already use their gig workers’ data to generate huge value. However, the value from the aggregate remains largely with the companies,  effectively offering only a trickle-down value to the workers. So, rather than ‘wrong’ financial incentives, such a structure is envisioned by the cooperative as redistributing the existing financial value of this data to the workers that generate it, to begin with. Further, by sharing with government agencies – workers are able to contribute to the larger value of their data to inform societal challenges, instead of receiving individual undervalued compensation from large companies.

On the matter of participation, for those running this cooperative, it was interesting to find that gig workers typically know their earning patterns are manipulated by algorithms – often resonating that “I have a theory about it, and how I can change the outcomes.” They want the digital tools to understand their data, and to gain more from it as a community of workers beyond what is afforded to them under large companies.

Who is the collective?’ : Decolonising data governance discourse

This Deep Dive also raised pertinent questions surrounding the language, power dynamics, and cultural nuances of data protection and management discourse. When we think about cooperatives as the ‘individual opting into a collective for common good,’ how are we defining this ‘good’? How many collectives are out there, who is defining the collective and where is the space for the marginalised, the more vulnerable within data envisioned as a collective? If one does not weigh these questions, how can one move toward a truly equitable data economy – one that does not end up appealing to systemic concepts that have harmed vulnerable communities.

A pure individualist vision of the data economy holds the risk of weaponisation against minorities or vulnerable communities. Notions of ‘greater good’ have historically ignored minorities, and it is important to protect these communities from opting into their own oppression through a possibly utilitarian conception of data governance. As such, Eurocentric technology and data governance solutions must be rethought in a way that empowers oppressed communities rather than adding another layer to existing disadvantages. Further, a lot of the marginalised members of a cooperative may have epistemic disadvantages even if they are part of a collective governance group. There is certainly a need to empower individuals in the data economy, but this restructuring must involve weaker sections as well – without oppressing them further or structurally infantilising their agency.

In closing, perhaps the need of the hour is larger than what may be considered ‘a group of people deciding for many others.’ Perhaps newer conceptions of democratic data governance must seek to empower not just ‘citizenry at large,’ but look to add complexity to the concept of the citizenry and ask who has agency within it. How can we use data and unlock its value to address a challenge as old as time – meaningfully bridging the gap for the vulnerable sections of our society?