The European Union has recently acquired enhanced authority to scrutinize artificial intelligence models, limit access to its market, and impose fines on providers. This development raises concerns for American firms such as Anthropic, OpenAI, and Google. Effective from Sunday, the EU Commission, which acts as the bloc's executive body, is empowered to evaluate AI models prior to their market introduction in Europe. They can also restrict market entry and impose penalties of up to 15 million euros or 3% of the company's annual turnover, depending on which amount is greater.
These new regulations could exacerbate tensions between the United States and the European Commission, especially as they navigate Europe’s ambitions for technological independence and the penalties that American tech firms may face. Notably, Google faced a significant $1 billion fine in July due to allegations of favoring its own services, prompting a response from U.S. President Donald Trump, who threatened to impose "substantial" tariffs on the EU.
This enforcement capacity aligns with the phased implementation of the EU AI Act, set to roll out in 2024. The EU AI Office is tasked with overseeing and enforcing regulations pertaining to general-purpose AI models during this transition. Henna Virkkunen, the executive vice president of the European Commission overseeing technology, security, and democracy, explained that improper design and use of AI pose potential dangers and that the most advanced models introduce unprecedented risks. Representatives from Anthropic, the White House, and the Department of Commerce have been contacted for their perspectives on these developments.
Given these regulatory changes, U.S. AI laboratories face heightened vulnerability as they navigate the complexities of compliance and the evolving landscape of international technology regulations.
