EU President Backs Frontier AI Slowdown, Plans Talks With AI Labs

European Commission President Ursula von der Leyen has backed recent calls from leaders of major U.S. AI companies to slow the development of advanced AI systems. Von der Leyen added that she will invite the companies for EU AI safety talks to discuss ways to “support ongoing industry efforts to pace the frontier.”

Why Did the EU President Call for AI Slowdown?

The European Commission President said the risks posed by self-improving models are increasingly becoming apparent. Leyen made reference to the OpenAI hacking incident in which AI agents escaped the testing environment and accessed Hugging Face systems.

"As the frontier models become more capable, these risks have come more sharply into focus. Models being developed will allow hacking on a level we never thought possible. And they will soon be in the hands of adversaries who see the world very differently from us," Leyen told EU lawmakers, as per Reuters.

Her position contrasts with that of U.S. President Donald Trump, who has previously dismissed warnings about the risks of slowing AI development. Earlier this month, the UN rights chief warned that AI could present an “existential threat to humanity” and called for an “all-out-effort” to minimize risks. 

Von der Leyen made the AI slowdown remarks while delivering her annual state of the union speech to the European Parliament in Strasbourg. 

What Makes Frontier AI Talks in the EU Critical? 

The recent incidents of rogue agents and warnings by researchers highlight the need to focus on AI safety. On September 11, Anthropic CEO Dario Amodei called on AI developers to slow the development of advanced AI capabilities. The request was supported by Elon Musk, who leads xAI, and OpenAI CEO Sam Altman.

"CEOs of the most advanced companies tell us that it is time to slow down on the self-recursive models. To pace the frontier. If the people developing the technology are clear, then we should be too," Leyen added. 

Von der Leyen also pointed to the EU AI Act, arguing that the bloc's regulatory framework gives Europe a role in shaping global efforts to manage AI risks. In addition to the AI labs meeting, the EU plans to collaborate with partners, such as Canada and the U.K., to strengthen AI safeguards through a range of measures. These include model testing, verification and security.

Leyen also proposed banning social media access for under-13 users, and restricting personal accounts for those younger than 15. The European Commission President argued that social platforms should demonstrate that they are safe for children. 

Impact of AI Risk Talks on the EU

The planned talks could give the EU a direct channel for discussing frontier AI risks with the companies developing the most advanced models. The initiative also signals a push for greater coordination between governments and AI companies on model testing, evaluation and security as AI capabilities continue to advance.