U.S. ban on Anthropic AI models shows world needs to 'build out and diversify'

Canadian Prime Minister Mark Carney used the U.S. export ban on Anthropic’s newest AI models, Mythos and Fable, to stress the dangers of global overdependence on a limited number of AI systems. Speaking during a visit to Ireland, Carney noted that while no wrongdoing occurred, the situation reveals systemic risks if nations fail to diversify their AI infrastructure. The ban, which blocks foreign access to Anthropic’s latest models, aligns with Carney’s broader policy focus on reducing Canada’s economic reliance on the U.S. He compared the AI risk to the 2008 financial crisis, warning that excessive concentration in AI tools could create vulnerabilities. ‘We have similar things in terms of model risk,’ he said, urging redundancy and diversification to mitigate potential failures. Carney confirmed that Canada and the U.S. maintain ‘a good flow of information’ on AI governance, though he acknowledged identified risks with Anthropic’s models. His remarks come as AI will be a key topic at the upcoming G7 summit in France, where leaders including U.S. President Emmanuel Macron will engage with executives from major AI firms, such as Anthropic’s Dario Amodei and OpenAI’s Sam Altman, at a scheduled lunch meeting. The prime minister cautioned that the G7 discussions will not produce immediate solutions but stressed the urgency of progress. ‘We need to make progress’ on AI, he said, adding that no definitive resolution will emerge from the summit. His comments reflect growing concerns over AI governance, trade dependencies, and the need for international collaboration to address emerging risks.