Anthropic, the San Francisco‑based AI lab, pulled its newest flagship models, Fable 5 and Mythos 5, from public use on Saturday night after receiving a direct request from the White House. The government ordered the company to block access for all foreign nationals, a restriction that also applied to Anthropic employees who are not U.S. citizens. In a brief statement, Anthropic said it "had little choice" but to comply, noting that the models were already subject to safeguards limiting deployment in high‑risk applications.
The shutdown unfolded with little warning. Within hours, developers, researchers and businesses that relied on the models found their accounts disabled. The abrupt loss of access highlighted the fragile supply chain for frontier AI tools and raised questions about the extent of U.S. governmental control over emerging technologies.
In London, AI and online safety minister Kanishka Narayan seized on the incident to argue for a stronger British AI ecosystem. Speaking at a press briefing, Narayan said the episode underscored the need for the United Kingdom to develop its own AI capacity, framing the issue as a matter of national security. "We treat every other threat to our sovereignty with deadly seriousness," she said, adding that AI is "the central political question of our time."
Across the Channel, former French prime minister Gabriel Attal labeled the shutdown the opening move of an "AI war" and warned that France's reliance on foreign AI providers could leave the nation vulnerable. Attal likened the restriction to a blockade of the Strait of Hormuz, suggesting that access to advanced models had become a strategic chokepoint. His remarks echoed broader European concerns about dependence on U.S. and Chinese technology, a theme that has shaped EU policy on chips, cloud services and AI for years.
Canada's finance minister, Mark Carney, echoed the call for diversification. He described the situation as a clear illustration of the risks inherent in over‑reliance on a single partner for critical AI resources. "Nobody has done anything wrong in the situation," Carney said, "but we will have done something wrong if we simply accept it and do not build out and diversify."
U.S. officials have not disclosed all the details behind the order, but reports indicate that the decision was partly driven by concerns that a group linked to China had accessed the Mythos model. Anthropic has previously accused Chinese rivals of using its models to train their own systems at industrial scale, a claim that adds a geopolitical layer to the shutdown.
The episode has reignited debate over AI sovereignty worldwide. France is already investing in home‑grown models such as Mistral, while Canada backs Cohere as a domestic alternative. Smaller nations like Singapore and the United Arab Emirates are focusing on niche applications and language‑specific capabilities. Open‑source initiatives also promise to democratize access to powerful models, though they have yet to match the scale of Anthropic's offerings.
Industry observers note that the shutdown, while disruptive, may accelerate efforts to reduce reliance on a handful of frontier labs based in the United States and China. The incident serves as a reminder that political decisions can swiftly alter the availability of cutting‑edge AI tools, prompting governments and companies to reassess their strategies for long‑term technological independence.
Anthropic has not indicated when—or if—Fable 5 and Mythos 5 will return to service. The company’s brief statement left the timeline ambiguous, and the broader AI community watches closely, hoping for a resolution that restores confidence in the stability of U.S.‑based AI infrastructure.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.