Discrimination and Inequalities in the Age of AI

Wednesday 30 April | 9h-17h
Sciences Po - Salle CS16
1, place Saint-Thomas 75007

Abstract

This workshop explores the ways artificial intelligence can contribute to biases, resulting in unfair treatment of specific groups, perpetuating or amplifying existing inequalities, or, conversely, playing a role in reducing bias and promoting fairness. The primary aim is to document and analyze AI-driven discrimination arising from factors such as biased datasets, flawed algorithms, and a lack of diversity in AI development teams.The workshop will examine the impact of AI-related biases across various domains. A key focus of the workshop is also to address solutions. Participants will engage in critical discussions on policy design, regulatory frameworks, and ethical considerations needed to mitigate these challenges. By fostering a multidisciplinary dialogue, the workshop aims to enhance our understanding of the relationship between AI and discrimination and provide strategies and recommendations to ensure AI systems are designed and implemented in ways that promote fairness, equity, and inclusion.

Coordination

Morgane Laouenan (LIEPP, CNRS) & Mirna Safi (CRIS, LIEPP)

Programme

9h15 | Welcome coffee

9h30 – 10h30 | Intrinsic and extrinsic biases in LLMs
Valerio Caprario (University of Milan-Bicocca)

10h30 – 10h45 | Coffee break

10h45 – 11h30 | Will Artificial Intelligence get in the way of achieving gender equality?
Daniel Carvajal (Aalto University)

11h30 – 12h15 | Generative AI & Labor Market Discrimination
Kenza Elass (Bocconi University and AXA Gender Lab)

12h15 – 13h30 | Lunch break

13h30 – 14h15 | How do migrants and natives search for jobs?
Roland Rathelot (CREST) & Marion Brouard (CREST)

14h15 – 15h00 | Political bias and diversity in recommendation algorithms' representations
Tim Faverjon (Sciences Po Medialab)

15h00 – 15h15 | Coffee break

15h15 – 16h30 | Roundtable Discussion :  Anti-discrimination in AI? the future of policies
Raphaele Xenidis (ScPo) & Benoit Rottembourg (INRIA)