EU countries and lawmakers fail to reach deal on weaker AI Act rules

EU countries and European Parliament lawmakers failed to reach agreement on a proposed overhaul of the bloc’s landmark artificial intelligence (AI) rules after 12 hours of negotiations on Tuesday, delaying efforts to simplify Europe’s tech regulatory framework.

The talks centred on revisions to the EU’s AI Act, which entered into force in August 2024 and is being phased in over the coming years. 

The changes are part of the European Commission’s wider “Digital Omnibus” initiative, which aims to streamline digital regulations including data protection and privacy laws in an effort to help European companies compete with rivals in the United States and Asia.

The AI Act is widely regarded as the world’s strictest artificial intelligence regulatory framework, introducing tougher requirements for so-called “high-risk” uses of AI such as biometric identification, healthcare, credit scoring, utilities management and law enforcement.

A Cypriot official, speaking on behalf of the rotating EU Council presidency, said agreement could not be reached with European Parliament negotiators and that talks would resume at a later date, likely within two weeks.

Lawmakers and member states reportedly clashed over exemptions for industries already covered by existing sector-specific regulations, including product safety rules. 

Some countries and lawmakers pushed for broader carve-outs, arguing duplication could stifle innovation.

Dutch lawmaker Kim van Sparrentak criticised the failure to reach a deal, warning that Big Tech firms would benefit from the deadlock while European companies seeking regulatory clarity face uncertainty.

The broader reform package also includes updates to GDPR, the e-Privacy Directive and the Data Act. 

Privacy and civil rights groups have raised concerns that the proposed changes could weaken protections and favour large technology companies.