By Claire Melamed, CEO, Global Partnership for Sustainable Development Data
READ: Salome Viljoen, Democratic Data: A relational theory for data governance. Available online at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3727562
This paper was only published in November and it’s already been recommended to me by about three different people. It’s a brilliant read — thorough, interesting, and absolutely relevant to the real-world policy issues that preoccupy many of us every day.
Data rights, and the dangers of digitization, are a big preoccupation in our network and beyond — it’s a hot topic for politicians right up to the UN Secretary General. This paper won’t reassure anyone, but it provides a very useful way of thinking about what the problem is and proposing solutions that are grounded in both economics and politics.
The main argument is that there’s something of a disconnect between what most worries us about big data and tech (potential for invasion of individual privacy, using people’s data to harm them through surveillance etc), and the primary economic driver of datafication, which is the value of population insights that come from classifying large numbers of people into groups and looking at common characteristics (Martin Tisne of Luminate made a similar argument in this paper, also published last year).
Data permeates our day to day lives through how it is used by companies and by governments to classify us all into groups, rather than the use of our single data points. This can harm us (think algorithmic injustice which imposes higher penalties on Black defendants, and Viljoen has a great example featuring a tattoo parlour), but also has huge benefits (think AI which diagnoses cancer).
This means that a focus on individual rights isn’t enough. I can stop using, say, WhatsApp because I’m worried about data sharing. But that won’t help me when data from tens of millions of WhatsApp users is aggregated and used to draw general conclusions about women my age who live in London — I’ll still be affected if those conclusions are used for policy decisions or by companies to deliver services.
The solution Viljoen proposes is greater focus on democratic oversight and governance of the uses of data. There are a number of more collective approaches already underway — data trusts, worker data collectives, and the data governance bodies that are increasingly a feature of democratic governments. This, rather than a sole preoccupation with individual rights and ownership of data, is the place to look for policy that fairly distributes the benefits and risks of datafication (of course it’s not either/or — democratic oversight and tough protection for individual rights are mutually reinforcing in lots of other areas of policy and law).
If I had one criticism, it would be that it’s very focused on the US regulatory regime and debates. A good companion read might be this really interesting comparison of different approaches to governing data around the world. But if you’re at all interested in data governance (and you should be!) then do read it.
Our Medium page serves as a space for a more informal conversation, to explore new angles and spark debate. For more Data4SDGs blogs, check out our website.
