US halts Anthropic's AI models over security concerns
The Trump administration has abruptly halted access to two of Anthropic’s latest AI models, forcing the company to disable Fable 5 and Mythos 5 just days after their launch. In a late-night notice, Anthropic confirmed it had received a directive from the US Commerce Department subjecting the models to export controls that restrict their use outside the United States. The company stated that compliance required immediately shutting down access for all customers, though other models remain unaffected.
A sudden regulatory move
The decision follows reports of a jailbreak that reportedly bypasses safeguards in Fable 5, allowing users to bypass restrictions on prompts related to cybersecurity, chemistry, and biology. An administration official, speaking to Axios, indicated that the pause was requested to give the national security apparatus time to address the vulnerability. The hardening of defenses could take “the next few weeks,” according to the source, suggesting a temporary but significant disruption for users of these models.
Balancing innovation and oversight
The episode highlights the growing tension between rapid AI development and government efforts to mitigate potential risks. While Anthropic emphasized compliance with the directive, the abrupt shutdown underscores how quickly regulatory actions can reshape the AI landscape. Industry observers will be watching closely to see how—or whether—Fable 5 and Mythos 5 return once security concerns are addressed.
Source: Ars Technica. AI-assisted editorial synthesis — TechnoExpress.