TOC4Fairness Seminar – Benjamin Fish

Date: Wednesday, March 1st, 2023
9:00 am – 10:00 am Pacific Time
12:00 pm – 1:00 pm Eastern Time

Location: Weekly Seminar, Zoom

Title: Relational equality and sociotechnical systems: towards a theory of discrimination in algorithms


Over the last decade or so, there has been a growing body of literature on discrimination and fairness in algorithms, especially machine learning. This has prompted the need for a theory of discriminatory algorithms: what properties can we guarantee?  To answer this question, any complete theory of discrimination (or unfairness, or bias, or injustice, etc.) in machine learning should be able to capture real discrimination, so that we call an algorithm discriminatory when it actually would be when implemented.  This requires a theory of discrimination that a) makes legible the normative consequences of algorithms, and b) captures the full scope of discrimination found in the real world, especially the most egregious and systematic forms of discrimination.  Unfortunately, our current theory often fails on both counts, rendering discrimination and unfairness narrowly as disparity in the outputs of algorithms, rather than how they are used and what consequences their use has.

This talk will discuss how we are failing right now, why it feels so hard to make progress, and a few threads of research that might help find a path forward. In particular, this talk will discuss relational equality as a meaningful alternative to create a more complete computational theory of discrimination.


Ben Fish is an Assistant Professor in Computer Science and Engineering at the University of Michigan. His research focuses on the foundations of fairness, justice, and other human values in machine learning and other computational systems. He was formerly a postdoctoral fellow at Mila and the Fairness, Accountability, Transparency, and Ethics (FATE) Group at Microsoft Research. He received his Ph.D. from the University of Illinois at Chicago as a member of the Mathematical Computer Science group. He was also previously a visiting researcher at the University of Melbourne and the University of Utah.