New EU enforcement powers over general-purpose AI models came into effect on Sunday, giving the European Commission authority to demand pre-release model evaluations, restrict EU market access, and fine providers up to 15 million euros or 3% of annual turnover, whichever is higher. The move raises the stakes for US labs including Anthropic, OpenAI and Google.
The powers are part of the staggered rollout of the 2024 EU AI Act and sit with the EU AI Office. They arrive amid friction between Washington and Brussels over tech sovereignty and fines on American companies. Google was hit with a 1 billion dollar fine in July after EU regulators found it gave preferential treatment to its own services, and US President Donald Trump subsequently threatened the EU with a substantial tariff.
Henna Virkkunen, European Commission executive vice-president for tech sovereignty, security and democracy, said in a statement that harms can occur if AI is not properly designed and used, and the most advanced models create risks on an entirely new scale.
The EU had sought access to Anthropic's Mythos model for months before the company agreed to share access. Reuters reported on Friday that the EU is in talks with OpenAI and Anthropic after recent cyber attacks by their models; OpenAI confirmed it was in contact with the EU AI Office, while the Commission and Anthropic were approached for comment. Anthropic, the White House and the US Commerce Department were also approached for comment.
Elisabetta Righini, a partner at Sidley Austin, said a US address does not put a lab outside the EU regulator's reach. Non-EU providers must appoint an EU-based authorised representative, and exposure to fines is real. She added that GPAI liability is not limited to substantive breaches: refusing an information request, giving misleading answers, or blocking a model evaluation is fineable on its own.
The Commission has previously acted on AI. In January, it opened an investigation into Elon Musk's X over sexually explicit material spread by the AI chatbot Grok. OpenAI's Tom Gordon, VP EMEA policy, said the company had collaborated closely with the Commission and ecosystem on implementing the AI Act and would continue working together to help Europe realise the benefits of the Intelligence Age. A Google spokesperson said Google remains dedicated to meeting all applicable rules as the Act and its codes take effect, with a primary mission of advancing European AI infrastructure and innovation.
Source: CNBC