The European Commission will decide whether to definitely scrap its planned liability rules for artificial intelligence systems by August, a Commission official told lawmakers in the European Parliament on Tuesday.

In the Commission’s 2025 work program, presented in February, the EU executive said it plans to scrap the AI Liability Directive because “no foreseeable agreement” is expected on the proposal.

The rules were intended to offer consumers a harmonised means of redress when they experience harm arising from AI products or services. They were proposed in 2022 but no significant progress has been made since. 

The Commission indicated, however, that the file could stay on the table if the EU Parliament and Council undertake to do extensive work on it over the coming year.

The Commission official told the Internal Market and Consumer Protection Committee (IMCO) on Tuesday that it is waiting for the official views of both the Parliament and the member states and will then “carefully think about the withdrawal”. 

The file has not been officially withdrawn, but the Commission need to do this in six months, starting from the publication of the work program. 

The Parliament itself is divided about the plans. 

The lawmaker responsible for steering the AI Liability proposal through the parliament, German MEP Axel Voss of the Legal Affairs committee, said the Commission’s move was a “strategic mistake”.

The rapporteur in the IMCO committee, Kosma Złotowski (Poland/ECR), said in his draft opinion published in January that the “adoption of an AI Liability Directive at this stage is premature and unnecessary.” The lawmakers in favour of withdrawing the rules say the consumers are protected by the Product Liability Rules as well as by the AI Act, which started to enter into force gradually.

EU Tech Commissioner Henna Virkkunen has been invited to the Legal Affairs committee on 9 April for a discussion on the topic. 

Member states have not yet discussed the proposal to get rid of the rules at working party level.