U.S. Government Forces Anthropic to Shut Down Fable 5 and Mythos 5: The Full Timeline, Global Fallout, and What Comes Next

The government's stated concern was a jailbreak vulnerability in Fable 5 that it assessed had national security implications. White House AI envoy David Sacks described the jailbreak as enabling the "operability of a cyber weapon" and argued that Anthropic had "prioritized the continued offering of the consumer model over safety". Anthropic disputes this characterisation, stating that the reported technique was "limited in scope" and "comparable to publicly known methods affecting other companies' models". The letter itself did not provide specific details of the national security concern.

Anthropic announced compliance within hours, disabling both models for all customers worldwide. The company said it "strongly disagrees" with the action and believes it is based on a misunderstanding, adding that it is "working as quickly as possible" to restore access. Notably, access to all other Anthropic models — including Opus, Sonnet, and Haiku — remained unaffected.

Global Repercussions: Allies Push Back

Japan, one of the closest U.S. allies and a country that had been actively integrating Anthropic's technology into its financial system, was among the first to respond. On June 20, the Japanese government released a draft revision of its Artificial Intelligence Basic Plan that explicitly references the U.S. export control action. Japan pledged to enhance cooperation with foreign government agencies and AI developers to address risks including AI-enabled cyberattacks and misinformation. Japanese analysts noted that the ban raises "AI sovereignty" concerns, particularly for Japan's financial sector and defense planning.

Japan's three megabanks — Mitsubishi UFJ, Sumitomo Mitsui, and Mizuho — had been on the verge of gaining access to Mythos 5, according to reporting by Nikkei, Reuters, and Bloomberg. The ban's inclusion of "all foreign nationals" meant that access by Japanese financial institutions was effectively cut off.

Broader unease spread among U.S. allies in Europe and the Asia-Pacific. The prospect of a unilateral "kill switch" on a commercially deployed frontier AI model raised fundamental questions about supply chain reliability for AI infrastructure and the long-term viability of depending on U.S.-controlled AI platforms. At the G7 summit in Evian, France, the issue reportedly became a topic of discussion among allied leaders.

Bipartisan Congressional Backlash

At home, the directive drew immediate skepticism from lawmakers. On June 18, a bipartisan group of House lawmakers — including Rep. Liccardo — sent a letter to the Trump administration demanding explanations for the sweeping restrictions and asking whether other U.S. AI companies should expect similar treatment. The lawmakers warned that the action "appears to represent a significant new application of export control authorities to advanced AI systems" and raises important questions for the broader U.S. AI ecosystem, American competitiveness, and the future development and deployment of frontier AI technologies. Several Hill Democrats expressed concern that the administration's decision may have been driven by other considerations.

Cybersecurity Experts Blast the Ban

Dozens of prominent cybersecurity figures publicly criticised the administration's directive in an open letter to Commerce Secretary Lutnick and National Cyber Director Sean Cairncross. Their core argument: cutting off foreign allies and researchers from frontier U.S. AI models will undermine American competitiveness, drive allied nations to develop or adopt rival Chinese AI systems, and deprive U.S. companies of the global talent and collaboration essential for AI safety research. The ban "has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it," the experts wrote.

Critics also noted the ban blocks the very researchers who could help fix the vulnerabilities the government claims to fear. More than 170 tech executives signed a separate open letter warning that the curbs "risked America's AI leadership" while China's capabilities advance rapidly.

Anthropic's Emergency Negotiations

Anthropic moved swiftly to respond. The company dispatched senior technical staff to Washington, D.C., for in-person meetings with White House officials starting over the weekend of June 13–14. The technical team had already been engaged in virtual discussions with administration officials since the directive was issued.

By June 18, Anthropic formally proposed a collaboration framework with the Trump administration to address the cybersecurity concerns while lifting the ban on Mythos 5 and Fable 5. The proposal includes joint security audits, enhanced guardrails, and transparency measures that would allow the models to be restored for foreign allies under monitored conditions. The discussions — spearheaded by Anthropic's public policy chief Sarah Heck and co-founder Tom Brown — aimed to create a shared framework of criteria that could be employed to evaluate future security breaches. Politico reported that the talks also indicate a recognition that no AI system can be entirely safeguarded against cyberattacks.

The negotiations represented the first instance in which a frontier AI lab has had to negotiate, in real time and at the highest technical level, for permission to keep selling its best product. As of June 20, 2026, the models remained suspended for all users, with no resolution announced.

A Historic Precedent

The export control directive marks the first time the U.S. government has directly compelled an AI company to revoke access to specific deployed model versions based on user nationality, rather than restricting hardware or underlying compute. Analysts from the Center for Strategic and International Studies noted that an approved export license from BIS is now required for any foreign person to access the models. Lawfare called it "a kill switch for frontier AI," arguing that in the long term, it would make it very difficult for Anthropic and other frontier AI companies to secure the foreign talent that is key to AI's development.

The episode has set a powerful and controversial precedent. Whether other AI companies will face similar treatment — and whether the U.S. can maintain the trust of its allies while exercising such sweeping control — remains an open question with profound implications for the future of AI governance.