Key facts
- European Commission President Ursula von der Leyen called for a discussion on AI safety and development pacing.
- Von der Leyen cited the EU's Artificial Intelligence Act as a potential foundation for global AI regulations.
- Top Silicon Valley bosses had previously called for a pause in AI development.
- Von der Leyen acknowledged that AI risks have sharply come into focus.
- The EU is seen as playing catch-up in frontier AI development compared to the U.S. and China.
European Commission President Ursula von der Leyen addressed growing concerns about artificial intelligence in her State of the European Union address, acknowledging the technology's potential threats and calling for a discussion on its development pace. While von der Leyen expressed optimism about AI's benefits, she also highlighted the sharp focus on its risks, including scenarios like AI models creating smarter versions of themselves, leading to an uncontrollable loop. She cited an incident where OpenAI-powered agents hacked another AI company as an example of emerging risks.
Von der Leyen proposed inviting AI frontier labs for a discussion on how to support industry efforts to pace development. She also suggested leveraging the EU's existing Artificial Intelligence Act as a foundation for global regulations and offered to collaborate with middle powers like the U.K. and Canada on testing AI models. However, her response to calls for a pause in AI development was primarily an invitation to dialogue, rather than a concrete action plan.
Critics noted that von der Leyen's approach, while acknowledging the issues, highlighted the limits of the EU's power in the global AI race. Experts like Max Reddel of the Arq Foundation questioned whether Europe could act with the speed and decisiveness required. Brando Benefei, an Italian MEP, pointed out that while Europe possesses regulatory expertise, it lags behind the U.S. and China in technological capacity. Daniel Privitera of the KIRA Centre AI think tank stated that more than just meetings are needed, urging Europe to do more to ensure a safe transition to transformative AI.