Fable 5 Shut Down by US Government

In an emergency Friday evening broadcast, the AI Daily Brief covered the unprecedented US government intervention ordering Anthropic to suspend access to its Fable 5 and Mythos 5 models for all foreign nationals, both inside and outside the United States. The directive, reportedly delivered by Commerce Secretary Howard Lutnick via letter to Anthropic CEO Dario Amadei, was triggered by a jailbreak report — allegedly produced by Amazon researchers — that demonstrated how a series of prompts could extract information about security vulnerabilities from the models. Anthropic pushed back forcefully, stating the jailbreak was narrow and non-universal, that the capabilities demonstrated were already available through other models including OpenAI's GPT-5.5, and that the government had not disclosed a specific harmful outcome resulting from the technique.

Anthopic's blog post revealed the directive arrived at 5:21 PM with no specific national security justification, and the company argued its defense-in-depth strategy — including a controversial 30-day data retention policy — was designed precisely to detect and mitigate such vulnerabilities. They contended that if this standard were applied industry-wide, it would effectively halt all new frontier model deployments. The Wall Street Journal reported that the government had previously asked Anthropic to pause releasing the models, and that Anthropic declined, with one source characterizing the response as Anthropic telling the government to 'pound sand.'

Criticism flowed in two main directions. Many condemned the government's action as arbitrary, legally questionable, and self-defeating — particularly given the administration's simultaneous posture of exporting advanced AI chips to China and allowing other frontier models to remain accessible. Commentators noted the hypocrisy of a White House that had recently stated government oversight of all AI models would have 'chilling effects on free speech and innovation.' The Department of War CIO's hostile tweet targeting Anthropic's 'revenue cycles' and 'pre-IPO valuation' led many to conclude the action was politically motivated rather than grounded in genuine security analysis.

However, an equally loud chorus directed blame at Anthropic itself, arguing the company's persistent safety scaremongering — comparing its models to nuclear weapons, advocating for government power to block model deployments, and publishing alarming public statements about existential risk — had created the regulatory and political conditions that made this intervention possible and perhaps inevitable. Critics pointed out the irony of Anthropic having explicitly called for government authority to block unsafe AI deployments, only to object when that authority was exercised against them. Viral commentary summarized the sentiment: Anthropic had named its model 'Fable' and then acted surprised when it came with a moral.

The downstream consequences discussed were sweeping. Practically, the directive means Anthropic cannot serve the models to foreign nationals even within the US, impacting a significant portion of its own technical workforce including high-profile researchers on EB-1 visas. Businesses using Fable 5 downstream — via API, through tools like Cursor or Harvey — face compliance nightmares without clear mechanisms for citizenship verification. OpenAI and Google DeepMind now face disincentives to release comparable models, since any jailbreak finding could trigger the same export control process against them. Anthropic's IPO prospects were described as severely damaged, with analysts questioning whether the company can justify frontier AI valuations if it cannot monetize its most powerful models globally.

Geopolitically, the action was seen as a watershed moment that validates the 'sovereign AI' movement, encourages European and other nations to hedge toward domestic or Chinese AI alternatives, and undermines the US narrative of being a reliable, rule-of-law technology provider versus China's arbitrary governance. Multiple commentators described this as the beginning of an intellectual 'iron curtain,' with access to frontier AI becoming a new axis of geopolitical power and inequality. The episode was compared to the 1990s cryptography battles and described as a precedent-setting moment that will define the trajectory of AI regulation for years to come.