The sudden suspension of Claude Fable 5 and Mythos 5 highlights the growing tensions between major AI innovation and national security oversight.
In a new development, it was reported that Anthropic has pulled back its most advanced AI models. The AI company has suspended access to its latest AI models, Claude Fable 5 and Mythos 5, after intervention fromUS national security authorities. The decision comes just days after the public rollout of the tools, which the company had previously described as its most powerful AI systems to date.
Giving an official statement, Anthropic said that it had been instructed to block foreign nationals from accessing Claude Fable 5. For complying, the company disabled both models entirely across its customer base, marking an unusual and abrupt reversal in a sector where product launches are often fast and expansive.
Why did the security agencies have to step in?
The core concern appears to center on “jailbreaking,” a process in which users bypass built-in software restrictions to unlock hidden capabilities or expose vulnerabilities. According to Anthropic, authorities flagged a method that could potentially exploit the model’s systems. While the company stated that the identified vulnerabilities were minor and already known internally, the incident has amplified fears about how advanced AI models can be manipulated for cyberattacks. Anthropic maintained that similar issues could also be found using other publicly available AI systems, but the sophistication of Fable 5 appears to have raised the stakes.
The suspension adds fuel to a larger conversation about AI governance. Before launch, Anthropic had openly acknowledged that Fable 5’s capabilities could be “dangerous” because of its ability to identify and exploit weaknesses in digital systems. That warning now seems more significant, especially as governments worldwide struggle to build policies around rapidly evolving AI tools. Experts say restricting access could slow down safe testing and international collaboration, but may also be necessary to prevent misuse.
Political pressure and legal battles intensify.
Anthropic is also facing mounting political scrutiny in Washington. Former President Donald Trump has publicly criticised the company, while the Pentagon recently labelled its tools a “supply chain risk”, a rare designation for a US-based technology firm.
The company is now challenging that label in court, underscoring a growing clash between Silicon Valley’s AI ambitions and government security priorities.
As AI capabilities accelerate, the Anthropic case may become a defining example of how far regulation can go in controlling next-generation technology.