Tim spoke at Data Connect 2024 on a panel, alongside Reema Patel (Digital Good Network/Public Voices in AI), Vinay Narayan (Aapti Institute) and Roshni Modhvadia (Ada Lovelace Institute), chaired by Octavia Field Reid (Ada Lovelace Institute)
Notes from Tim’s presentation are included below. A recording of the session will be available on the Data Connect 2024 platform for registered participants.
Speaking notes
I’ve got five key points I want to briefly share with you for thinking about participatory and inclusive data stewardship: 2 preparatory, 2 practical, and one cautionary.
Point 1: Trust is not a tickbox.
So often, when it comes to data governance, systems and processes rely on individual consent.
“Tick the box to confirm you’ve read the terms of service”
and so on.
Of course, we know almost no-one reads the ToS, and even if they did, when we’re providing public services people often do not have the choice not to proceed past checking the box.
So, the important - perhaps obvious - point to start with, is that we can’t assume that because people are using services and sharing data with them, that they trust them.
Nor, do I think, headline surveys about trust, whilst useful as a broad barometer, tell us all that much about how publics feel about the way particular datasets or resources are being stewarded.
And it is important to note that we want not only public trust that data will be well managed and not misused, but also trust that public data resources will be used well for the public good.
Point 2: It’s not just about data subjects.
When we consider participatory stewardship of data, we might first thing about data subjects - the people in the dataset - and stewarding for their interests, and engaging them as individuals in participatory processes.
But it’s important to remember that data can have collective impacts: and the use of data may impact on many people who are not data subjects.
In a recent paper with MyData, Datasphere Initiative and Aapti Institute we described this as a distinction between “Data from me”, “data about me” and “data that affects me”.
For public sector organisations that may emerge as three questions:
- “Who is this data from?” - and what rights do they have over it?
- “Who is this data about” - and how might we engage with them? and
- “Who does this data affect?” - and do we understand those effects and impacts?
Point 3: Identify those affected
Which brings us to the first practical point: identify those affected.
This can be easier said than done: because whilst the first step of meaningful participation is to invite the right people to the process, in the cases of our interlinked data systems, we don’t always have immediate visibility on all the people and communities our data stewardship decisions affect.
For example, combining local administrative data to be more proactive in offering social services or outreach support, might lead to certain communities opting out of accessing important services altogether because of fears about how their data might be used - or there may be communities who are underrepresented in the administrative data and so receive less proactive support than they should.
Spending a few hours using consequence scanning methods, searching the academic literature, or even, if all you have are a few minutes, turning to ChatGPT or Co-Pilot to ask something like “What kinds of communities or population groups might our data project impact upon?” can be useful starting points to shape the design of participatory stewardship activities.
Point 4. Provide different pathways of participation
When I started at Connected by Data, one of my first tasks was to look at case studies of participatory data governance from around the world. What was striking about the most effective was that they never consist of one component: they have a range of elements - combining for example, open opportunities to comment, with deep-dive workshops with particular communities, with collective processes such as roundtable meetings, citizens juries or public representation on oversight boards to weigh up what they have learned and link publics and decision makers.
Participatory data stewardship is not about involving everyone all the time. But it is about always being open to new participants becoming involved, and providing pathways to involvement that can help increase the diversity of viewpoints and perspectives at the table when decisions are being made.
I can talk in discussion if it is relevant about our experience at Connected by Data with citizens jury models around AI governance: in which we used random selection, and then deliberative processes, to bring together citizens with no background on AI and equip them be actively involved in AI governance debates. Or we can touch upon how the Mind Kind project by Wellcome trust involved young people in workshops, oversight groups and as embedded team members when designing data governance for a mental health app. Or I can point you to our draft set of playing cards that include over 30 different mix-and-match methods for participatory stewardship.
The important point here also is that participation is not a one-off activity, but something ongoing and iterative to embed within stewardship practice.
Point 5: If the answer is always more data, the question is wrong
The final point I want to make is about meaningful participation. I’ve seen a lot of processes that are reasonably engaging, inform participants well about the issues, collate in-depth analysis of public views… and then their conclusion is to collect and use data in ways that were planned all along.
Meaningful participation should lead to change.
And in some cases, that change may be less data, or different data.
It’s important to be clear in any process what it is within your power to change: and to recognise that sometimes the issues raised by publics are not solved by data, but might be solved by other parts of the government system.
Given data is about just about everything - to be a participatory data stewards may also mean being ready to share views onwards, or signpost participants onwards, to other parts of government that address the needs raised.
Join the community
It was great to have over 80 data professionals from across government with us for this Data Connect session, and to be able to explore questions about sustainability and securing buy-in, engaging marginalised communities, and linking data and AI governance.
If you want to dig into these conversations more, you might be interested in our Data and AI Public Engagement Community of Practice.