EU AI Act Enforcement Powers Take Effect, OpenAI, Anthropic, and Other AI Companies Face Stricter Regulation

August 3. The European Commission obtained new AI regulatory enforcement powers on August 2, which allow it to conduct model assessments, restrict market access, and impose fines on companies developing General Purpose AI (GPAI) models. U.S. AI companies such as OpenAI, Anthropic, and Google are key targets of this enforcement.

According to the EU's "Artificial Intelligence Act" (EU AI Act) implementation arrangements, the European Commission may require companies to submit assessments of their models before entering the European market, and impose fines of up to €15 million or 3% of the company's global annual revenue (whichever is higher) for violations.

This new power is seen as a significant measure for the EU to promote "technological sovereignty" and may further intensify EU-U.S. tech regulatory friction. Previously, the EU fined Google €1 billion under the Digital Markets Act (DMA), and former U.S. President Trump threatened to impose "massive" tariffs on the EU.

The EU's AI Office stated that the risks posed by advanced AI models are increasing, necessitating enhanced regulation. Analysts point out that even AI companies headquartered in the U.S. cannot evade EU regulation; non-EU AI service providers must also appoint an EU-based authorized representative as a regulatory contact point.

Recently, the EU has intensified communication with companies such as OpenAI and Anthropic, focusing on AI model-related security risks. OpenAI confirmed it is collaborating with the EU's AI Office.

OpenAI's Vice President of Policy for Europe, the Middle East, and Africa, Tom Gordon, stated that the company will continue to collaborate with the European Commission and industry ecosystem to drive AI legislation and promote AI development in Europe.