BIAS 2026

Bias and fairness in AI systems are central challenges in modern AI and are more relevant than ever. The rapidly evolving AI landscape, including the rise of Generative AI models, increasingly agentic systems, and the widespread deployment of AI-powered applications, has significantly increased both the societal impact of AI and the amplification and emergence of new risks.

While early work on fair ML and AI bias focused primarily on the technical aspects of discriminatory algorithms and treated fairness as an objective to be optimized, this view is no longer sufficient. Bias and fairness are inherently multidimensional, extending beyond technical considerations to legal, ethical, and societal dimensions. At the same time, the technical nature of the problem is changing: bias emerges across the entire AI pipeline and can be amplified in agentic settings where systems act autonomously in dynamic environments. Addressing these challenges requires both new technical approaches and the involvement of diverse stakeholders in the design and evaluation of AI systems.

Now in its 6th edition, this workshop has become an established forum within the ECML PKDD community for advancing research on bias and fairness in AI. It brings together researchers and practitioners to address these challenges, spanning both the evolving technical nature of bias in modern AI systems and its broader societal dimensions.