The European Commission officially gained new authority on Sunday to investigate, fine, and compel changes from AI model developers as the bloc's Artificial Intelligence Act moves toward full implementation. This regulatory shift is expected to heighten tensions between the United States and the European Union, which have long clashed over Europe's push for technological sovereignty and its frequent penalties against American tech firms. In July, after the EU fined Google $1 billion under the Digital Markets Act, U.S. President Donald Trump threatened to impose "massive" tariffs on the EU.
Under the freshly enacted enforcement powers, the EU now has the legal authority to review AI models, restrict access to the European market, and penalize model service providers—significantly raising compliance risks for U.S. firms such as Anthropic, OpenAI, and Google. These powers took effect on Sunday. As the bloc's executive body, the European Commission can require companies to undergo assessments before releasing models to the EU public, block market entry, and impose fines of up to €15 million or 3% of a company's global annual turnover, whichever is higher.
The new powers may ignite further transatlantic friction. Europe's ongoing pursuit of technological autonomy, combined with its repeated fines on U.S. tech companies, has deepened divisions. In July, EU regulators found Google to have favored its own services, slapping it with a $1 billion penalty. Trump responded by threatening high tariffs on the EU.
Where these powers fit in the AI Act's rollout
The EU's Artificial Intelligence Office's oversight and enforcement authority over general-purpose AI models is part of the phased implementation of the 2024 Artificial Intelligence Act, which includes transition periods for certain provisions. Henna Virkkunen, the European Commission's Executive Vice President for Tech Sovereignty, Security, and Democracy, stated: "If AI is designed and used improperly, it can cause various harms; the most advanced frontier models can bring unprecedented systemic risks." Journalists have contacted Anthropic, the White House, and the U.S. Department of Commerce for comment.
Why U.S. AI labs are the primary targets
The rapid acceleration of AI capabilities in recent months has further strained relations between major AI labs and the EU. The bloc has spent months seeking access to Anthropic's Mythos model, until the company eventually agreed to open it up. As U.S.-EU government relations tighten, Europe is doubling down on building technological sovereignty to reduce its reliance on U.S. AI systems. Reuters reported last Friday that the EU is in discussions with OpenAI and Anthropic because their models could be used to launch cyberattacks. OpenAI has confirmed it is in communication with the EU's Artificial Intelligence Office. Journalists have sought responses from the European Commission and Anthropic regarding the report.
Elisabetta Righini, a partner at the law firm Shearman & Sterling, stated in an interview that these enforcement powers apply to any company offering a general-purpose AI model within the EU, regardless of its headquarters location. "Even if a company is based in the U.S., it cannot escape EU regulatory jurisdiction. Non-EU service providers must also appoint an authorized representative in the EU to serve as a contact point for regulators," Righini said. She added that the risk of fines for AI companies is not hypothetical. "What is often overlooked is that penalties related to general-purpose AI are not limited to substantive violations. Refusing to provide information, giving misleading responses, or obstructing model evaluations can all trigger fines."
Although this new enforcement is rooted in the Artificial Intelligence Act, the EU has previously taken regulatory action on AI. In January, the bloc announced an investigation into Elon Musk's X platform over its AI chatbot, Grok, for disseminating pornographic content. Tom Gordon, OpenAI's Vice President of Policy for Europe, the Middle East, and Africa, told the new rules: "We are working closely with the European Commission and the entire industry ecosystem to implement the Artificial Intelligence Act, including the accompanying code of conduct. We will continue to collaborate to help Europe seize the opportunities of the intelligent age." A Google spokesperson stated: "As the Artificial Intelligence Act and its code of conduct take effect, we will fully comply with all applicable regulations, upholding our core mission: to advance Europe's AI infrastructure and innovation."