EU: European Parliament publishes study on AI and civil liability

The study recommends revising the AI Liability Directive to introduce a strict liability regime for high-risk AI systems.

On July 24, 2025, the European Parliament published a study titled ‘Artificial Intelligence and Civil Liability: A European Perspective,’ commissioned by the European Parliament’s Policy Department for Justice, Civil Liberties, and Institutional Affairs at the request of the Committee on Legal Affairs. The study analyzes the EU’s approach to regulating civil liability for artificial intelligence (AI) systems.

What does the study analyze?

The study looked at the inadequacy of existing legal frameworks: The Product Liability Directive 1985 (PLD), the Product Liability Directive 2024 (PLDr), and the AI Liability Directive (AILD). In particular, the study outlines that the AILD, as it stands, is an inadequate legal framework and gives multiple reasons. These reasons include that the AILD relies solely on procedural mechanisms, which are insufficient to address the challenges AI presents, including its complexity, opacity, and autonomy. Additionally, the AILD’s attempts to integrate national fault-based regimes while introducing new concepts such as fault and causation can lead to legal uncertainty, risking conflicting interpretations across Member States.

Furthermore, the study explores different policy options:

  • option 1: Withdrawing the AILD, which would lead to fragmentation;
  • option 2: Maintaining the AILD without substantive structural modifications;
  • option 3: Maintaining the AILD and introducing a fault-based liability rule for high-risk AI systems; and
  • option 4: Maintaining the AILD and introducing a strict liability regime.

What does the study recommend?

The study advocates for revising the AILD to introduce a strict liability regime targeting high-risk systems. According to the study, a strict liability regime would:

  • ensure clear, predictable compensation for victims;
  • minimize litigation and administrative costs;
  • enable risk internalization via insurance and pricing mechanisms; and
  • maximize harmonization across the EU.

You can read the study here.