The European AI Liability Directives -- Critique of a Half-Hearted Approach and Lessons for the Future

The EU's proposed AI liability framework relies too heavily on disclosure requirements and narrow legal presumptions rather than substantive rules about how AI systems should be developed and deployed, representing what the paper calls a "half-hearted approach" that may inadequately protect affected individuals.

A critical gap exists between the EU AI Act and the liability proposals: the AI Act lacks individual rights for people harmed by AI systems, while the liability directives lack specific technical and operational requirements for AI development, leaving enforcement mechanisms incomplete.

The paper identifies that current liability proposals focus on mechanisms like requiring companies to disclose evidence and establishing limited presumptions about fault and defectiveness, which the author suggests are insufficient given the unique risks AI systems pose compared to traditional products.

The EU's regulatory approach may trigger a "Brussels Effect," where European standards become globally influential, potentially shaping AI liability frameworks in the United States and other regions—making the adequacy of these proposals consequential far beyond Europe.

The paper proposes amendments and outlines a roadmap for more comprehensive future AI liability and regulation frameworks that would better address key risks posed by AI systems, though the abstract does not detail the specific recommendations.