Statement on the US government directive to suspend access to Fable 5 and Mythos 5

2026-06-12 · Source: Anthropic News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy · Depth: Intermediate, short

Summary

The US government, citing national security authorities, issued an export control directive on June 12, 2026, mandating the suspension of access to Anthropic's Fable 5 and Mythos 5 for all foreign nationals, including Anthropic employees. This directive, received at 5:21pm ET, forced Anthropic to disable these models for all customers globally to ensure compliance. While the government did not provide specific details, Anthropic understands the concern relates to a method of "jailbreaking" Fable 5. Anthropic reviewed the technique, finding it identifies minor, known vulnerabilities also discoverable by other public models like OpenAI's GPT-5.5. Anthropic maintains its "defense in depth" strategy for Fable 5's safeguards, which were extensively red-teamed and found more effective than previous models, with no universal jailbreak discovered. They disagree with the broad recall based on a narrow, non-universal jailbreak, arguing it sets an unworkable industry standard. Access to all other Anthropic models will not be affected.

Key takeaway

For AI Security Engineers evaluating model deployment risks, this incident highlights that even narrow, non-universal jailbreaks can trigger significant regulatory action, despite robust "defense in depth" strategies. You should proactively benchmark your model's vulnerabilities against publicly available alternatives and prepare for potential government directives, advocating for transparent, technically grounded policy processes.

Key insights

A narrow AI model jailbreak prompted a US government directive, forcing Anthropic to suspend Fable 5 and Mythos 5 access globally.

Principles

• Perfect jailbreak resistance is unlikely.
• Defense in depth mitigates AI risks.
• Government oversight needs transparency.

Method

Anthropic employed a "defense in depth" strategy for Fable 5, combining strong safeguards, extensive red-teaming, and 30-day customer data retention for monitoring and mitigation.

In practice

• Red-team AI models for thousands of hours.
• Implement data retention for vulnerability research.
• Compare model vulnerabilities to public alternatives.

Topics

• AI Model Safeguards
• Export Controls
• National Security
• AI Jailbreaking
• Fable 5
• Mythos 5
• Anthropic

Best for: CTO, VP of Engineering/Data, Executive, Policy Maker, AI Security Engineer, Director of AI/ML

Related on AIssential

• Statement on the US government directive to suspend access to Fable 5 and Mythos 5 — The US government ordered Anthropic to suspend Fable 5 and Mythos 5 access due to a "jailbreak" vulnerability, citing national security.
• Democrats try to tie corruption to affordability — Anthropic released Fable 5, a powerful AI with safety guardrails, to balance advanced capabilities with public security.
• US government forces Anthropic to disable Claude Fable 5 and Mythos 5 for all customers worldwide — Government-mandated global disabling of Anthropic's Fable 5 and Mythos 5 due to a suspected jailbreak sets a contentious industry precedent.
• Breaking news: US Commerce Department effectively shuts down Anthropic’s latest models — The US Commerce Department has imposed export controls on Anthropic's Fable 5 and Mythos 5 models, restricting foreign national access.
• US limits use of Anthropic AI models Fable 5 and Mythos — Government AI safety concerns are escalating, leading to direct restrictions on frontier model access and development.
• Anthropic blocks all public access to Claude Fable 5, Mythos 5 following US government order — what enterprises should do — Centralized frontier AI models face abrupt government intervention risks, demanding enterprise diversification and redundancy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Anthropic News.