Health Data and the Politics of Classification

Biobanks, electronic health records systems, and other large health data infrastructures are proliferating, driving transformations in biomedical research and practice. Beyond the clinic, these datasets are also increasingly seen as the solution for other sorts of problems, and are deployed as training data to develop machine learning and artificial intelligence models and algorithmic tools used in insurance markets, in the management of labor, and even in surveillance and policing.

In this talk, I argue that it is not simply the newly-available quantity of health data, but the relentlessly particular qualities of individual datasets that have been most impactful — and which raise trenchant concerns for social equity and justice. As I will show, decisions about whose data is collected and how they are measured have cyclically reshaped our understanding of human diversity along a number of dimensions. In doing so, these infrastructures are changing our understanding of clinical classifications, and resourcing new ML/AI tools which seek to classify people and populations for other purposes. I make the case that health data represent a unique challenge for both extant data ethics and bioethics frameworks, demanding urgent attention as they are increasingly used to train socially consequential ML/AI technologies in and beyond the clinic.

This lecture will be held both online & in person. You are welcome to join us either in South Hall or via Zoom.

Join the lecture online