European Commission withdraws AI Liability Directive from consideration

The IAPP reports that the European Commission has decided not to revive its draft AI Liability Directive, saying there is no foreseeable agreement among those involved, while the technology sector presses for simpler rules. The choice is recorded in the work program for 2025, adopted on 11 February and presented to the European Parliament a day later.

The proposal dates from 2022, which was two years ahead of the finalization of the AI Act, and it was intended to set uniform rules for some parts of non-contractual civil liability for damage involving AI systems. In an annex to the work program, the Commission said it will assess whether to put forward another proposal or choose a different approach.

German member of the European Parliament Axel Voss criticized the move, arguing that the directive would have worked after harm occurred, whereas the AI Act aims at prevention. He attributed the reversal to industry lobbying and said the Commission had given in. Lawyer Peter Craddock of Keller and Heckman said dropping a proposal is preferable to withdrawing a law after passage, though he pointed to the question of how harmonized liability will now be pursued. Victims of AI-related discrimination are not left without recourse, he said, but national regimes matter, and outcomes rarely differ greatly.

The article places the decision within a drive to simplify digital regulation. Commission President Ursula von der Leyen said at the Paris AI summit that the Commission must cut red tape. Practitioner Otto Lindholm welcomed simplification, citing a tangle of overlapping laws. The industry group CCIA Europe welcomed the shelving but wants a wider review of planned digital legislation. Compliance specialist Ani Nozadze warned that deregulation can produce uncertainty and uneven national rules, raising compliance costs, especially for smaller firms.