Topic: Making AI “ethical” - What is data solidarity, and why do we need it?
The lecture will be presented by Prof Barbara Prainsack, of the Department of Political Science at the University of Vienna.
The challenge of developing effective policies and guidelines for the ethical governance of AI is a much-discussed issue. Studies show that ethical guidelines for AI in particular have had little impact on practice so far. In this talk, I will argue that a root cause of this problem is the blindness of most policies and guidelines to the relational nature of data. Data are not atomistic bits of information, but they have emerged from, and encapsulate, relations between different kinds of entities. Digital data also disclose information about wider ranges of entities than those that the data come from. Digital practices such as predictive analytics have further deepened the relationality of data. Today, the relationality of data is the de facto business model underpinning entire industries, such as targeted advertising or social media. Because the intrinsic relationality of data is not, however, sufficiently recognised in the way data are conceptualised and governed, corporations and authorities can pretend that the data they collect can be separated from people, while profiting from their relationality at the same time. I will end by outlining “data solidarity” as an approach that acknowledges the relationality of data, and with it, the stakes that people - as individuals and as members of collectives - have in phenomena in the world that are datafied. In such a way, data solidarity is a precondition for the ethical governance of AI.
Date: Tuesday, 18 February 2025
Time: 12:30 - 14:00 SAST
Venue: The Oliver Tambo Moot Court,
Faculty of Law, UCT Middle Campus/ Online via MS Teams
Live lecture link to be shared shortly.
