Inherent limitations to AI fairness

by Prof. Tijl De Bie (University Ghent, Belgium)

Tuesday 9 Jul 2024, 10:00 → 11:30 Europe/Berlin

OH14/E04 (Fakultät Informatik)

Abstract - In this talk I will discuss our recent CACM paper with the same title (https://dl.acm.org/doi/10.1145/3624700), on a number of inherent limitations of purely technical approaches to AI fairness. In doing so, I will argue for the need for a socio-technical approach to tackling AI fairness issues. I will also briefly survey some other work in our group on the topic of AI fairness.

Short bio - Prof. Tijl De Bie is a senior full professor at Ghent University. Before joining Ghent University in 2015, he studied or held research positions at KU Leuven, UC Berkeley, UC Davis, Southampton University, and the University of Bristol. He has worked on the foundations of machine learning and data science, as well as on applications of AI in fields ranging from bioinformatics, over music informatics, sports analytics, to social and mainstream media analysis. His current research interests include exploratory data science, trustworthiness of AI, the impact of AI on information integrity and democracy, and applications of AI to the labor market and human resources management. His work has been funded by several prestigious research grants, including an ERC Consolidator grant (FORSIED), an ERC Proof of Concept grant (FEAST), and an ERC Advanced grant (VIGILIA) due to start in October this year.

Publication: Inherent Limitations of AI Fairness

Tijl De Bie's Webpage

Zoom Link