BIAS at the Applied Machine Learning Days AMLD 2024

02.04.2024 Members of the BIAS project and the Generative AI Lab at BFH organised a track on the topic of fairness and bias in AI applications for the labour market at the Applied Machine Learning Days AMLD in Lausanne. With success.

The Applied Machine Learning Days AMLD is a global platform that brings together experts and participants from over 40 countries across industry, academia, and government. The conference is composed of different tracks, addressing dedicated topics. In this year’s edition, members of the BIAS project organised a track around the topic Fairness and Bias in AI Applications for the Labor Market in collaboration with NLP expert Elena Nazarenko from Lucerne University of Applied Sciences and the Generative AI Lab from the Bern University of Applied Sciences. 
The track was initiated by a short presentation of the BIAS project by Mascha Kurpicz-Briki, the leader of the technical work package. The EU Horizon project brings together an interdisciplinary consortium of nine partner institutions to develop a deep understanding of the use of AI in the employment sector and to detect and mitigate unfairness in AI-driven recruitment tools.

Connecting to this introduction, Eduard Fosch-Villaronga from BIAS partner Leiden University illustrated how designing AI solutions profoundly impacts society and needs a socio-technical participatory effort incorporating technical, social, economic, political, and legal considerations. He showcased how the different activities in work package 2 supported the BIAS project. For example, the literature review helped the project to look at previous efforts at a theoretical level. The mapping exercise and the expert interviews helped to understand the state of the art and the thinking of professionals working on these areas. Finally, the survey helped the project to map the worker and public attitudes and worries towards automation in the labor market. At the end of the presentation, Eduard encouraged the audience to join the National Labs, a pool of diverse stakeholders, including employers, workers, HR practitioners, AI developers, trade union representatives, civil-based society organisations, and scholars, whose professional expertise and/or experience could contribute to the BIAS project.

The session was followed by a talk from Preethi Lahoti from Google Research, discussing AI Safety and Fairness in Large Language Models. As the last speaker of the first session, Alejandro Jesús Castañeira Rodriguez from the company presented the technical approaches behind their job matching software. After a well-deserved coffee break, Christoph Heitz from the Zurich University of Applied Sciences (ZHAW) navigated between philosophy and technical solutionism, discussing how to address the socio-technical nature of fairness and bias in theory and practice. Cynthia Liem from the Delft University of Technology shared experiences and findings of interdisciplinary research collaborations bridging research in the field of AI and the labor market. Finally, two short talks from Jana Mareckova from the University of St. Gallen and Pencho Yordanov from the Adecco Group concluded the session.

Each talk was followed by a Q&A session, and the interested audience engaged actively with the speakers, enabling a rich discussion.

Find out more