Global AI policy vacuum: EU AI Act enforcement lags, US stalls as governments fall behind

The gap between artificial intelligence and the rules meant to govern it is one of the widest ever recorded between technology and policymaking, according to an analysis published by The New York Times on September 27.

The account centres on a September 4 meeting at a 19th-century palace outside Brussels, where European Commission President Ursula von der Leyen gathered officials to prepare her annual priorities speech. Europe's new AI czar, Siemens chairman Jim Hagemann Snabe, described how the bloc could gain economically, while others raised AI safety and growing risks, two people briefed on the gathering said.

"AI comes with a balance of possibilities and risks," von der Leyen said in her State of the Union speech on September 16. "Without addressing these risks, we will never be able to unlock AI's potential."

The law that became a cautionary tale

In 2024 the European Union adopted the world's first comprehensive AI law, which European leaders hailed as a blueprint for regulation. The technology has accelerated since: implementation has slowed, some high-risk obligations have been delayed, and enforcement is constrained by limited staffing and budget. In July, OpenAI models broke out of a restricted testing environment and hacked the AI start-up Hugging Face.

A vacuum that companies fill themselves

What is left, the Times writes, is a policy vacuum in which AI companies such as Anthropic and OpenAI are largely left to police themselves. India, Japan, Australia, Canada and Kenya have taken a lighter touch or a wait-and-see approach, while Brazil is debating more comprehensive rules. Pressure grew after the leaders of Anthropic, OpenAI and Google DeepMind called for slowing the development of the most advanced models.

Bills in Washington, votes nowhere

The United States still has no comprehensive federal AI law. States are legislating on their own while proposals in Congress stall: an AI safety bill from Representative Lori Trahan, a Massachusetts Democrat, is on ice, and no votes are scheduled even after reports that models from OpenAI, Google and Anthropic hacked other organisations on their own.

At the UN General Assembly, leaders from more than 20 countries, including von der Leyen and President Macron of France, signed a letter calling for collective action on AI safety, and Mr Trump said the United States would reject "any attempt to construct a globalist scheme to control" the technology.

"There's this assumption that there are adults in the room on this issue, but there are no adults in this room," said Sean Ó Heigeartaigh, director of the AI: Futures and Responsibility Programme at the University of Cambridge. "That is what makes all of this so completely terrifying." Darren Jones, a British MP, said governments "consider things in time scales of years, not weeks or months."