Anthropic CEO urges G7 leaders not to ‘splinter’ on AI

Anthropic CEO Urges G7 Leaders to “Resist the Temptation to Splinter” on AI

Anthropic CEO Dario Amodei addressed G7 leaders on Wednesday at a working lunch in Évian-les-Bains, France, calling on the world’s wealthiest democracies to maintain unity on artificial intelligence development rather than fracture into competing blocs. The remarks, first reported by the Financial Times, came on the final day of a summit that placed AI governance at the center of geopolitical tensions between the United States and its allies.ft

The plea arrived against the backdrop of an unprecedented U.S. Commerce Department directive issued the previous week that barred all foreign nationals — including Anthropic’s own non-citizen employees — from accessing the company’s most advanced models, Fable 5 and Mythos 5. Anthropic disabled the models globally to ensure compliance, calling the government’s action unwarranted.cnet

AI Takes Center Stage at G7

The lunch session, titled “Ensuring a safe, rapid and effective deployment of artificial intelligence,” drew a roster of tech leaders including OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis, and Mistral AI’s Arthur Mensch, alongside G7 heads of state including President Donald Trump and French President Emmanuel Macron.youtube

Altman used the forum to call for an “international forum” where nations could establish guidelines for AI safety and create universally accepted standards for testing. Macron, meanwhile, turned the Anthropic episode into a diplomatic argument. He said the Trump administration’s restriction on Anthropic’s models had “clarified the stakes” for the United States and its allies, according to the Financial Times. While acknowledging it was “a good thing” that Washington recognized frontier AI models could be dangerous, Macron called the response “strictly nationalist” and “a bad thing”.usnews

A Fraught Week for Anthropic

The company found itself at the center of an international dispute after the Commerce Department invoked export control authority on June 12 to block foreign access to Fable 5 and Mythos 5, citing an unspecified national security concern believed to relate to a potential jailbreak of Fable 5’s safety guardrails. More than 50 cybersecurity professionals publicly called on the U.S. government to reverse the ban, while Anthropic employees reportedly complained that the administration had “bullied” the firm.infosecurity-magazine

Broader Implications

Macron used his closing press conference to argue that regulation of AI “has now become imperative,” warning that neither political nor business leaders could ignore the technology’s impact on democracies. European officials at the summit signaled a desire to turn the crisis into an opportunity for partnership rather than confrontation with Washington. Whether the episode marks a turning point or a temporary friction depends largely on whether the U.S. lifts the restrictions — a question that remained unanswered as the summit concluded Wednesday evening.politico