Government Intervention Halts Anthropic's Latest AI
In a swift and unprecedented move, the U.S. government has ordered Anthropic to suspend access to its newly launched advanced AI models, Fable 5 and Mythos 5. The directive, issued on June 12, 2026, cites national security authorities and specifically prohibits access by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. To comply with the order, Anthropic has "abruptly disabled" both models for all its customers globally.
The timing of this intervention is particularly striking, coming just three days after Anthropic publicly launched Fable 5. Mythos 5, the more powerful and less-safeguarded version of the model, was initially deployed through Project Glasswing in collaboration with the U.S. government and was intended for a broader trusted access program soon. This marks the first instance in AI history where a commercial large model deployed to hundreds of millions of people has been forcibly recalled by the U.S. government.
The "Jailbreak" Controversy and Anthropic's Disagreement
While the government's directive did not provide specific details of its national security concern, Anthropic's understanding is that the order stems from the belief that a method of bypassing, or "jailbreaking," Fable 5 has been discovered. A "jailbreak" refers to a technique that can circumvent the AI model's safety controls. Anthropic, however, strongly disagrees with the government's assessment and the resulting recall.
In a public statement, Anthropic argued that the identified "jailbreak" technique revealed a "small number of previously known, minor vulnerabilities" that can also be found in other publicly available models without requiring a bypass. The company emphasized that if such a standard were applied across the industry, it would effectively halt all new model deployments for all frontier model providers. Anthropic stated its belief that the government should have the ability to block unsafe deployments, but only through a statutory process that is transparent, fair, clear, and grounded in technical facts, a standard they believe this action did not meet.
A History of Safety Warnings and Strained Relations
This incident highlights a complex and increasingly tense relationship between advanced AI developers and government oversight. Anthropic has been a vocal proponent of AI safety, even building in layered restrictions and requiring 30-day data retention for Mythos-class models to detect and mitigate potential misuse. The company had also worked with the U.S. government, the UK AISI, and other organizations for thousands of hours to "red-team" Fable 5's safeguards prior to its launch. These tests, according to Anthropic, showed that Fable 5's safeguards were substantially more effective than those of any previously deployed model, with no testers finding a "universal jailbreak" that could broadly bypass its protections.
However, the government's action follows a period of strained relations. Earlier this year, Anthropic's relationship with the government reportedly "ruptured" after it refused to allow the U.S. military to use its AI models for domestic surveillance and fully autonomous weapons systems. This led to Anthropic being placed on a supply chain blacklist, set to take effect later in the year. The current export control directive, issued by Commerce Secretary Howard Lutnick, represents a significant escalation in U.S. efforts to control access to advanced AI capabilities, shifting the focus from chips and tools to restricting access to the AI models themselves.