Mark Carney says Anthropic ban shows risk of relying on big AI models

Canadian Prime Minister Mark Carney criticized the US export ban on Anthropic’s latest AI models, Mythos and Fable, during a visit to Ireland on Sunday. He described the situation as a cautionary example of systemic risks arising from over-reliance on a small number of powerful AI tools, echoing his broader policy focus on diversification. Carney linked the AI issue to Canada’s economic dependence on the US, particularly under former President Donald Trump’s tariffs on key industries like autos and steel. He also noted US pressure on Canada to reverse digital sector taxes affecting Silicon Valley firms, framing AI governance as part of a broader geopolitical and economic strategy. The prime minister compared the risks of AI concentration to the 2008 financial crisis, where interconnected banks exposed systemic vulnerabilities. He argued that redundancy and diversity in AI development are necessary to mitigate potential failures, stating that ‘we will have done something wrong if we just accept this, don’t take the lesson, don’t build out and diversify.’ AI will be a key topic at the upcoming G7 summit in France, where Carney has already discussed the issue with French President Emmanuel Macron. The meeting will include leaders from major AI firms, such as Anthropic’s CEO Dario Amodei and OpenAI’s CEO Sam Altman, for a dedicated lunch session on Wednesday. Carney acknowledged the need for progress but cautioned that the G7 summit would not produce a definitive solution, emphasizing the complexity of the challenge. He also noted a ‘good flow of information’ between Canadian and US governments on AI risks, though concerns about Anthropic’s latest models remain. Carney’s warnings reflect a broader push for international collaboration to prevent over-centralization in AI development.