21 January 2022

Discovering Data Trusts with the Data Science Group

The first Data Science Talk of 2022 took place on the 14 January and zoomed in on the topic of Data Trusts. The data environment has power imbalances and this creates vulnerabilities for various groups of people and individuals who can be targeted in a way that creates both, social division, and discrimination. Various data governance frameworks, such as Data Trusts, can help to remedy this issue, as well as empower the beneficiaries. This webinar provided the opportunity for members of the Data Science Group, and participants from foundations, to learn more on data trusts, the role they can play, and the issues faced with their implementation from two expert speakers, Sylvie Delacroix and Chloé Vandendriesche.

Sylvie Delacroix, Co-Founder, Data Trusts Initiative, introduced the topic by highlighting the origin of a “trust”, which can be traced back to the 11th and 12th century in the UK. She began with the powerful statement that power that stems from aggregated data to individuals and groups can indeed be returned. She continued by explaining that by pooling data rights, not necessarily the data itself, Data Trusts can solve three problems: 1. A lack of tools that enable long-term collective action; 2. That data consent is rarely more than ‘make believe’; 3. A lack of governance that removes obstacles to the research potential underlying datasets. In addition, Data Trustees have a fiduciary obligation which is overseen by a court with significant intervention powers.

Sylvie went on to discuss some of the implementation choices of when a Data Trust is the chosen governance framework. This includes whether the governance structure is highly participatory or completely hands-off. Secondly, the data does not need to be pulled at all, rather it is the task of the Data Trustee to exercise the data rights on behalf of the beneficiaries. Considering the challenges that may arise, she noted that there can be a variety of interests which in turn leads to a conflict of trust. The other challenge relates to a habit of passivity that has been encouraged by current data governance, including dismissing data privacy pop ups that appear when entering a website. She finished her presentation by inviting participants to read the Ada Lovelace Institute’s report on “Legal mechanism for data stewardship.”

On presenting a concrete example, Data Scientist, Chloé Vandendriesche, King Baudouin Foundation, presented the DataTrust pilot project which has been pioneered by the social profit sector in Belgium. Recognising the slightly different definition of what might be a Data Trust in Belgium compared to the definition laid out by Sylvie, she highlighted the same aim of wanting to have a place where data can be shared and that this shared data is put back in the hands of people, and that trustful institutions which has this data are used. In essence, Chloé explained that the platform is essentially a Knowledge Centre based on data from the social profit sector in Belgium, which is managed by various stakeholders. Going into the technical aspects, she stated that it is built using the Windows Azure programme. Data has been collected from official public institutions in Belgium and the software organises this big data in a way that is easy to use. Explaining the governance structure, she explained the various stakeholders involved which places the platform in the hands of the people of the sector. She went on to explain the technical aspects of the platform, along with providing a live demonstration of how it worked.


Daniel Spiers
Programme Officer, Peer-Exchanges & Knowledge