started · updated
Ursula von der Leyen calls for controlled slowdown of frontier AI development
In her State of the Union address, European Commission President Ursula von der Leyen called for a more responsible approach to artificial intelligence, emphasizing the need to manage the risks of frontier models. She specifically advocated for a controlled slowdown in the development of self-improving, auto-recursive models, noting that major AI industry leaders have expressed similar concerns.
Von der Leyen highlighted significant security risks, warning that advanced AI could enable hacking at unprecedented levels and could fall into the hands of adversaries. To mitigate these threats, she announced plans to collaborate with partners including Canada and the United Kingdom on AI model evaluation, verification, and security. She also stressed that AI should empower professionals rather than replace them, advocating for a 'European model' that maintains human intervention in critical sectors like healthcare.
However, some critics, such as independent MEP Michael McNamara, have questioned the EU's ability to act as a global regulator. McNamara argued that because the most influential AI developers are based outside of Europe, the EU's regulatory ambitions may amount to little more than rhetoric, especially given the lack of dominant European AI models.
Entities
European Commission · European Parliament · European Union · Hugging Face · Michael McNamara · Ursula von der Leyen
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 2 SOURCES] Ursula von der Leyen warned that advanced AI models could enable unprecedented levels of hacking. cybernoz.com · the-decoder.com
- [○ 1 SOURCE] The EU aims for a model of AI development that emphasizes human intervention and professional empowerment rather than replacement. elnacional.com.py
- [● 2 SOURCES] The European Commission President supports a controlled slowdown in the development of self-improving frontier AI models. cybernoz.com · wwwhatsnew.com
- [● 2 SOURCES] The EU plans to collaborate with Canada and the United Kingdom on AI model evaluation, verification, and security. cybernoz.com · the-decoder.com
- [● 2 SOURCES] MEP Michael McNamara characterized the EU's ability to regulate frontier AI as mere rhetoric due to the lack of domestic dominant models. de.euronews.com · es.euronews.com