On 28 September 2022, the European Commission proposed a directive on adapting non-contractual civil liability rules to artificial intelligence, commonly called the AI Liability Directive. It was designed to complement the AI Act: while the Act sets safety and governance requirements, the liability proposal addressed what happens when an AI system causes harm and a person seeks compensation.
The proposal aimed to lay down uniform rules for certain aspects of non-contractual civil liability for damage caused with the involvement of AI systems. Its central purpose was to address the specific difficulties of proof that arise with AI, where the complexity and opacity of systems can make it hard for an injured party to show what went wrong. The Commission’s measures, including a rebuttable presumption of causation and rights to disclosure of evidence about high-risk systems, were meant to ensure that justified claims are not blocked by these evidentiary obstacles. The proposal responded to a 2020 European Parliament resolution and to commitments in the Commission’s White Paper on AI.
The directive’s path was uneven; the Commission later signaled in early 2025 that it would not pursue the proposal further, leaving AI-specific liability rules unsettled in the EU. The episode matters because it highlights a hard, often-overlooked question in AI policy: when an automated system causes injury or loss, who is responsible and how can a victim prove it. That question remains live for businesses deploying AI even where formal liability rules have stalled.