European Commission President Ursula von der Leyen has called for a more cautious, staged rollout of frontier artificial intelligence models, arguing that the pace of development must not outstrip the ability of regulators and societies to assess risks. Speaking on Tuesday morning, she framed the issue as a test of Europe's capacity to lead in both innovation and governance.
A call for measured progress
Von der Leyen's remarks come as AI labs in the United States and China push ahead with increasingly powerful systems. She acknowledged the economic potential of these technologies but insisted that "speed without guardrails is not progress." The Commission President emphasised that the EU's forthcoming AI Act, already among the world's most comprehensive regulatory frameworks, should be complemented by dynamic oversight that adapts as models evolve.
Her proposal envisions a system where frontier models—those with capabilities approaching or exceeding human performance in certain tasks—are released in phases. Each stage would require rigorous testing, transparency about training data, and clear accountability for downstream uses. "We need to know what these systems can do before they are everywhere," she said, addressing an audience of tech executives and policymakers in Brussels.
The European approach contrasts with the more permissive stance taken in some other jurisdictions. In Washington, the Trump administration has been criticised for its handling of AI, with some commentators pointing to a White House out of touch with the technology's real-world implications. Von der Leyen did not name any country directly, but her emphasis on European values was unmistakable.
Industry reaction has been mixed. Some European startups worry that excessive caution could stifle innovation and drive talent abroad. Others, however, see an opportunity to build trust. "Regulation can be a competitive advantage if it is predictable and science-based," said a spokesperson for a Berlin-based AI firm, speaking on condition of anonymity.
The Commission President also linked the AI debate to broader geopolitical tensions. With Ukraine seeking more military support, including air defence systems, ahead of a possible meeting between President Zelenskyy and former President Trump, von der Leyen noted that AI could play a role in both defensive and offensive capabilities. "We must ensure that these tools are used in line with international law," she said.
Her call for pacing is not entirely new. Earlier this year, a group of AI researchers and ethicists published an open letter urging governments to adopt similar measures. But von der Leyen's intervention gives the idea institutional weight, and it is likely to shape the EU's negotiating position in international forums.
The European Parliament has already shown appetite for strict rules. Several MEPs have pushed for bans on certain high-risk applications, such as real-time facial recognition in public spaces. The Commission's latest proposal could provide a middle ground, allowing innovation while maintaining oversight.
As the morning news cycle unfolds, the story is developing. For a broader look at today's key stories across the continent, see the September 11 briefing. And for a deeper dive into how AI is reshaping education in conflict zones, read about children's learning gaps in war-affected regions.
Von der Leyen's statement is expected to be followed by a formal proposal in the coming weeks. The debate over frontier AI is far from settled, but Europe is positioning itself as a key arbiter in the global conversation.
