EU lawmakers renew push for AI liability law after Altman comments
The European Commission withdrew its proposal amid a drive to cut regulatory burdens, saying it lacked sufficient political backing.
Members of the European Parliament are pressing for new AI liability legislation after OpenAI chief executive Sam Altman said companies would need a framework governing their responsibility. His comments have strengthened lawmakers’ calls to reconsider a proposal withdrawn by the European Commission.
Altman said OpenAI would have to bear some responsibility if problems arose with its models during training. German conservative lawmaker Axel Voss, who led Parliament’s work on the proposal, said the remarks supported bringing liability legislation back onto the EU agenda.
The Commission first proposed the law in 2022, but later abandoned it as part of efforts to reduce companies’ regulatory burden. It said the measure was unlikely to attract enough political support. Voss argued that it would have addressed responsibility during training and testing, before models were deployed.
Social democrat Brando Benifei said withdrawing the proposal was a mistake, arguing that the EU’s AI Act was intended to operate alongside liability rules. He supports strict liability for frontier models and high-risk systems. Benifei, Irish independent Michael McNamara and Greens member Kim van Sparrentak are seeking legislation to help people and European businesses obtain redress when providers cause harm.
EU tech chief Henna Virkkunen said she would examine possible gaps in existing legislation, while stressing caution over new rules. Artur Bogucki, research director at the think tank CEPS, called for a narrowly focused liability measure that would include requirements to record and disclose evidence.
