Government orders Anthropic to shut down its most powerful AI models over security fears

BitcoinWorld

Government orders Anthropic to shut down its most powerful AI models over security fears

The U.S. government on Friday ordered Anthropic to immediately disable access to two of its most advanced AI models — Claude Fable 5 and Claude Mythos 5 — citing national security concerns. Anthropic confirmed it complied with the directive, received at 5:21 pm ET, but made clear it believes the government overreacted. The order blocks access for all users worldwide, not just foreign nationals targeted by export controls. Other Anthropic models remain unaffected.

Why Mythos and Fable 5 drew government attention

Mythos is Anthropic’s most capable AI model, previewed in early April and kept under tight restrictions due to its exceptional ability to find security vulnerabilities in software. According to Anthropic, Mythos identified flaws in every major operating system and web browser it tested. Rather than release it broadly, the company launched Project Glasswing, a controlled program sharing Mythos with roughly 50 vetted organizations, including Amazon, Apple, Google, Microsoft, and CrowdStrike, for defensive cybersecurity work.

Fable 5, released just three days before the shutdown, was designed as a safer, publicly accessible version of Mythos. Anthropic added guardrails to block responses in high-risk areas like cybersecurity and biology, arguing it was safe for general release. Benchmark tests from Vals AI showed Fable 5 immediately became the most capable publicly available AI model.

The government’s case and Anthropic’s rebuttal

The government framed its directive as an export control action, but Anthropic said in a blog post that the underlying concern appears to be a claimed jailbreak of Fable 5. The company stated that the government has provided only verbal evidence of a potential narrow, non-universal jailbreak — essentially prompting the model to read a specific codebase and identify software flaws. Anthropic noted that this level of capability is already widely available in other publicly accessible models, including OpenAI’s GPT-5.5, and is routinely used by cybersecurity professionals for defensive purposes.

Anthropic argued that its strongest safeguards operate through independent classifier systems separate from the model itself, meaning that even if someone convinces Fable to keep talking past a refusal, the underlying protections remain in place. The company also stated that a review of recent usage found no evidence of those safeguards being successfully bypassed to produce harmful content.

A clash of safety strategies

Anthropic has built much of its public identity around being the safety-conscious alternative to rivals like OpenAI. The irony is that the very caution Anthropic displayed in restricting Mythos — promoting it as a model so dangerous it couldn’t be released publicly — appears to have attracted the government scrutiny that now threatens its business. OpenAI CEO Sam Altman had previously criticized Anthropic’s approach, calling it ‘fear-based marketing’ and warning that when you spend months telling the world your AI is uniquely dangerous, the world tends to listen.

What this means for the AI industry

Anthropic is widely expected to pursue an IPO this year, and this regulatory intervention could complicate those plans. The company warned that if the government’s standard for recalling a model — a narrow potential jailbreak — were applied across the industry, it would essentially halt all new model deployments for all frontier AI providers. The incident highlights the growing tension between AI safety practices, government regulation, and commercial ambitions in the rapidly evolving AI landscape.

Conclusion

The shutdown of Claude Fable 5 and Mythos 5 marks a significant escalation in government oversight of frontier AI models. It raises fundamental questions about how safety claims, export controls, and national security concerns will shape the future of AI development and deployment. For now, Anthropic has complied with the order while publicly challenging its rationale, setting the stage for a broader debate on AI regulation.

FAQs

Q1: Why did the U.S. government order Anthropic to shut down these models?
The government cited national security concerns, framing the action as an export control measure. However, Anthropic believes the real reason is a claimed narrow jailbreak of Fable 5 that the company disputes.

Q2: Are other Anthropic models affected?
No. The order only applies to Claude Fable 5 and Claude Mythos 5. Other Anthropic models remain available.

Q3: What is Project Glasswing?
Project Glasswing is a controlled program through which Anthropic shared its most powerful model, Mythos, with about 50 vetted organizations for defensive cybersecurity work. It was not part of the public shutdown order.

This post Government orders Anthropic to shut down its most powerful AI models over security fears first appeared on BitcoinWorld.