“The idea of imposing civil and criminal liability on model developers when downstream users do something bad is so misguided and such a dangerous precedent. First off, the bill requires developers to prove that their models can’t possibly be used for any of the defined hazardous capabilities.”
EUEuropean Parliament & Council of the EU (ordinary legislative procedure)
AI Liability Directive (COM(2022) 496)
Proposed 28 Sep 2022 to ease compensation claims for AI-caused harm via a rebuttable presumption of causality and evidence-disclosure duties. Parliament's IMCO committee called adoption 'premature and unnecessary' (May 2025 opinion); JURI work was still under way when the Commission pulled the file. BEUC, CDT Europe, Mozilla and others urged the Commission to begin work on replacement AI liability rules (7 Apr 2025).
Withdrawn. Commission announced withdrawal in its 2025 work programme (11 Feb 2025), citing 'no foreseeable agreement'; formally withdrawn via OJ C/2025/5423 of 6 October 2025. No replacement proposal as of July 2026.as of 2025-10-06primary source ↗
Vehicles carry asks; asks outlive them
Asks on this vehicle
- Wants more rulesImpose a developer duty of care for catastrophic harmsSuperseded
- Wants more rulesAdopt replacement EU AI liability rulesActive
Sorted by what filers wanted done to the rules
Claims filed here
“While well-intentioned, SB 1047 does not take into account whether an AI system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data. Instead, the bill applies stringent standards to even the most basic functions — so long as a large system deploys it.”
“The EU needs rules to address the legal gaps left by the AILD withdrawal and ensure a fairer, simpler path to compensation for all people affected by AI systems, including consumers in the EU, in case of harm by an AI system.”
Joint letter with ARTICLE 19, ECNL, CDT Europe, Enforce and Mozilla.