Anthropic announced it is pulling access to its latest large‑language models, Mythos 5 and Fable 5, after receiving a directive from the White House. The order, delivered at 5:21 p.m. Eastern time, instructs the company to block the models for "any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." Anthropic said the directive stems from U.S. government worries about a "potential jailbreak" that could compromise the models' safety controls.

The company clarified that the restriction applies only to Mythos 5 and Fable 5; all other Anthropic models continue to operate normally. Because the rule cannot be applied selectively at this time, Anthropic has suspended the models globally. In a blog post, the firm apologized for the disruption and said it is working to restore access as quickly as possible, describing the situation as a misunderstanding of the actual risk.

Mythos 5 represents Anthropic's next‑generation AI system, while Fable 5 is the more publicly available version of the same technology, offered to trusted partners. Both models were marketed as significant upgrades in coding, reasoning, and agentic capabilities, and they underwent a thorough vetting process before release. Anthropic highlighted that Fable 5’s safeguards are "substantially more effective than those of any previously deployed model," yet the government’s concerns appear to outweigh those assurances.

According to Anthropic, the company collaborated closely with authorities in the United States and the United Kingdom, as well as multiple third‑party agencies and internal teams, to ensure the models met safety standards. Despite those efforts, the White House directive suggests that the existing safeguards are not deemed sufficient to address the identified jailbreak risk, which the company characterizes as "narrow" and "non‑universal."

Customers reacted with a mix of frustration and disbelief. One user called the situation an "absolute nightmare scenario," while another dismissed it as a joke. The abrupt removal of access has left developers who had recently integrated Fable 5 into their workflows scrambling for alternatives. Anthropic’s statement that the issue is a misunderstanding has done little to calm concerns about the reliability of future model deployments.

Industry observers note that Anthropic’s experience could signal broader challenges for frontier AI providers. In its statement, the company warned that if the same standard were applied across the sector, it could effectively halt the rollout of new, high‑performance models. The episode underscores the growing tension between rapid AI innovation and governmental efforts to mitigate national‑security risks.

Anthropic did not disclose the specific nature of the jailbreak or the technical details that prompted the order. The company said it remains in close contact with U.S. officials and is prepared to reinstate access once the concerns are resolved. For now, developers and enterprises awaiting the capabilities of Mythos 5 and Fable 5 must pause their plans while the dispute unfolds.

Questo articolo è stato scritto con l'assistenza dell'IA.
News Factory APP - notizie agentiche per potenziare il tuo SEO e AEO.