Anthropic export ban sparks G7 'trusted partners' talks

The Trump administration has blocked Anthropic from exporting its newest Mythos 5 and Fable 5 models on national security grounds. This regulatory action followed concerns flagged by Amazon to the White House, warning that certain safety guardrails could be bypassed. The export ban, which was discussed during the G7 (Group of Seven, an intergovernmental political forum) summit on Wednesday, has highlighted the vulnerability of non-U.S. entities that rely on American technology, raising fears that access to critical software could be cut off without warning.

During the summit, world leaders expressed growing anxiety that the U.S. could cut off access to top American AI models at any time. French President Emmanuel Macron warned that the U.S. “from one day to the next can turn off the switch” on access, which could harm European customers and AI firms alike. Indian Prime Minister Narendra Modi also voiced concerns regarding the restrictions, stating that democratic nations must have unfettered access to top AI models to protect critical infrastructure. In response to these risks, G7 leaders discussed a ‘trusted partners’ scheme. This proposed framework would grant non-U.S. nations and trusted companies access to advanced AI models from firms like Anthropic and OpenAI, establishing an open trade network to bypass unilateral U.S. export restrictions. The discussions highlight a growing consensus among international allies that relying solely on U.S. regulatory decisions poses a significant risk to their own technological infrastructure.

The export ban has also drawn criticism from the private sector. Aidan Gomez, the co-founder and CEO of enterprise AI firm Cohere, stated that the recent restriction on access to Anthropic’s models confirms that relying on a small handful of big tech companies is dangerous to the resilience of companies and democratic nations. Gomez emphasized that digital sovereignty is about who controls the foundational technology shaping economic security and national sovereignty. The situation leaves startups in hubs like Paris and Bangalore facing the reality that their products could break overnight if U.S. access is suddenly revoked. This reliance on U.S. AI infrastructure has forced international companies and governments to reckon with the possibility that access can be revoked overnight.

Why it matters

The U.S. government’s willingness to restrict AI exports on national security grounds creates a ‘sovereignty risk’ for international companies and governments, forcing a rethink of reliance on American AI infrastructure.