The future of AI liability in the European Union

Details

Supervisors

Strowel, Alain

Faculty

Faculté de droit et de criminologie

Degree label

Master [120] en droit, à finalité spécialisée: Droit européen

Abstract

enRecently, AI has become a central feature of modern economies and societies. Its integration into a plethora of sectors raises not only new opportunities but also complex legal and ethical challenges. Among these, the question of liability for harm caused by AI systems has emerged as a critical issue. In this evolving legal landscape, the EU has taken several initiatives to address these challenges. Alongside the AI Act, two instruments have been particularly central to the debate on AI liability: the revised Product Liability Directive (rPLD), adopted to modernise the 1985 framework, and the AI Liability Directive (AILD), a proposal intended to complement the rPLD by covering fault-based, non-contractual civil liability for AI-induced harms. While the rPLD was successfully adopted, the European Commission proposed to withdraw the AILD in early 2025, leaving notable gaps and uncertainty. This thesis aims to highlight how the rPLD adapted to digital technologies – particularly AI. While it represents a major milestone in the field of liability, it still does not cover the entirety of damages that could be caused by AI systems, a loophole that was intended to be addressed by the AILD. Eventually, this research attempts to evaluate the consequences of this withdrawal and provide guidance for future regulatory proposals in the field of AI liability.