export control directive
In plain wordsA US government order, used against AI models in June 2026, telling a company to stop foreigners from using some of its AI models for national-security reasons.
AI policy (United States)
An order issued by the US government under export-control and national-security authorities to a single company, requiring it to restrict access to a specific item by foreign nationals. On 12 June 2026 AnthropicAnthropic, prop. n.: An American AI company that makes the Claude chatbot and models. It was founded in 2021 by former OpenAI staff and puts particular emphasis on AI safety research. received such a directive, reported as coming from the Department of Commerce: it required the company to suspend access to Claude Fable 5 and Claude Mythos 5 by any foreign national, inside or outside the United States, including its own staff. The government said it believed it had found a way to jailbreakjailbreak, n.: A trick for getting an AI chatbot to break its own rules — for example, to give dangerous instructions it would normally refuse — usually through cleverly worded requests. Fable 5; Anthropic said the demonstration it reviewed had surfaced only a few previously known, minor vulnerabilities. To comply, Anthropic disabled both models for all customers. The directive was lifted at the end of June 2026 and access was restored.
Legal teams now ask whether a launch could trigger an export control directive before the weekend.
Origin
The phrase comes from Anthropic's statement of 12 June 2026 that it had received such a directive; a law-firm alert described the episode as the first known US use of export-control powers to regulate a particular frontier model on national-security grounds.
Usage note
Unlike export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI. or the rescinded AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy., which set general rules, the directive targeted one company's models through an individual order rather than rulemaking. Anthropic complied but publicly disagreed, arguing that a narrow potential jailbreak was no ground for recalling a commercial model and that such action needs a transparent statutory process.
Related
- export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI.
- AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy.
- Project GlasswingProject Glasswing, prop. n.: An Anthropic program, launched in 2026, that gave a hacking-capable AI model only to organizations defending important software instead of releasing it to everyone.
- jailbreakjailbreak, n.: A trick for getting an AI chatbot to break its own rules — for example, to give dangerous instructions it would normally refuse — usually through cleverly worded requests.
- Center for AI Standards and InnovationCenter for AI Standards and Innovation, prop. n.: The US government's AI testing and standards center, part of the National Institute of Standards and Technology. It evaluates advanced AI models, American and foreign, for national-security risks and works with companies on voluntary testing.
- dual-usedual-use, adj.: Describes technology or knowledge that can be used both to help and to harm: the same AI that finds software bugs for defenders can find them for attackers.
- AI-assisted vulnerability discoveryAI-assisted vulnerability discovery, n.: Using AI to find security holes in software — and often to suggest fixes — faster than human experts alone, so that defenders, or attackers, can act on them.
Sources
- Anthropic (2026). Statement on the US government directive to suspend access to Fable 5 and Mythos 5. 12 June 2026. anthropic.com/news/fable-mythos-access
- Greenberg Traurig (2026). AI Company Anthropic Suspends Access to Claude Fable 5, Claude Mythos 5 Following US Export Control Directive. June 2026. gtlaw.com/es/insights/2026/6/ai-company-anthropic-suspends-access-to-claude-fable-5-claude-mythos-5-following-us-export-control-directive
- GovInfoSecurity (2026). US Lifts Export Curbs on Anthropic AI Models. govinfosecurity.com/us-lifts-export-curbs-on-anthropic-ai-models-a-32123
Mentioned in
export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI., Center for AI Standards and InnovationCenter for AI Standards and Innovation, prop. n.: The US government's AI testing and standards center, part of the National Institute of Standards and Technology. It evaluates advanced AI models, American and foreign, for national-security risks and works with companies on voluntary testing., AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy., Project GlasswingProject Glasswing, prop. n.: An Anthropic program, launched in 2026, that gave a hacking-capable AI model only to organizations defending important software instead of releasing it to everyone.