Anthropic Shuts Down AI Models Following Government Order Over Security Concerns

Anthropic CEO Dario Amodei has consistently issued warnings regarding the potential threats posed by the company's advanced AI models. Amodei specifically cited Mythos, and later Mythos 5 and Fable 5, as models that could present serious risks to cybersecurity, the financial sector, critical infrastructure, and national security.

In an essay published this month, Amodei described AI's power as "undeniable" and highlighted the "very real risks" associated with models like Mythos. He advocated for more robust government intervention to address these concerns.

On Friday, the US government intervened, ordering Anthropic to block foreign access to its Mythos 5 and Fable 5 models. In compliance, Anthropic responded by shutting down both models completely.

The Pentagon's chief information officer expressed support for the government's action in an X post. The development prompted varied reactions within the tech industry. AI researcher Gary Marcus labeled the Trump administration's move as "wildly overdramatic and also counterproductive." Conversely, Yann LeCunn, an AI pioneer, attributed the situation to Amodei's prior warnings, stating on X, "Dario Amodei's ridiculous fear mongering about Mythos/Fable (and AI in general) finally pays off. One reaps what one sows."

Amodei's history includes leaving OpenAI to co-found Anthropic, reportedly due to concerns that OpenAI prioritized product releases over safety. He has since continued to articulate potential risks from AI, suggesting it could disrupt life, including the elimination of many entry-level white-collar jobs and a rise in unemployment.

In his recent essay, Amodei cautioned that cyber risks from "Mythos-class models" would not be the last, predicting that "biological risks may soon follow, and that serious AI autonomy risks may not be far behind." Anthropic also recently called for a temporary halt to frontier AI development, arguing that advanced models are approaching self-improvement, which could lead to humans losing control over AI systems.

According to Business Insider, this requested pause arrived on Friday, at least for Anthropic.