U.S. Export Controls Trigger Worldwide AI Model Shutdown

Empty courtroom with wooden tables, chairs, and a door.

The U.S. government just forced one of America’s top AI firms to shut off its most powerful models worldwide, all in the name of “export controls” that many say went too far, too fast.

Story Snapshot

U.S. export-control officials ordered Anthropic to block foreign nationals from using its new Fable 5 and Mythos 5 AI models over national security fears.
Because Anthropic says it cannot reliably check every user’s nationality, it disabled both models for everyone, including Americans.
The Trump Commerce Department points to a possible “jailbreak” that could help hackers find software flaws, but has not shared full details.
Anthropic argues the government misunderstood the risk and is fighting to restore access while complying with the law.

What The Government Ordered Anthropic To Do

The U.S. Department of Commerce used export-control powers to order Anthropic to suspend access to its newest artificial intelligence models, Fable 5 and Mythos 5, for all foreign nationals, including those inside the United States, citing national security concerns.^[5] Anthropic says the directive arrived late on a Friday and came with no detailed explanation of the threat, only a clear legal demand to cut off foreign access.^[5] Reports say Commerce Secretary Howard Lutnick sent a formal letter to Anthropic chief executive officer Dario Amodei explaining that the models now fall under strict export rules, similar to those used for advanced chips or military technology.^[2]Because the order covers “foreign persons” anywhere, not just users overseas, Anthropic must treat every cross-border request as a possible violation unless it can prove the user is an American citizen.^[4] That framing shows how Washington now views top-tier artificial intelligence: as a strategic asset that adversaries could weaponize, not just another consumer app.^[3] This marks one of the first times U.S. export controls have targeted specific artificial intelligence models instead of only hardware like semiconductors, setting a strong signal for how future systems may be handled.^[3]

How A Security “Jailbreak” Triggered A Global Shutdown

Anthropic says it understands the government’s concern is about a possible “jailbreak” of Fable 5 that could bypass built-in safety rules and let users identify software vulnerabilities at scale.^[1] In simple terms, officials fear that hostile actors could use the model like an automated lock-picker for computer systems, helping them find and exploit weak points faster than human hackers.^[5] Anthropic’s own launch materials had highlighted that Fable 5 and Mythos 5 can run complex tasks for long periods without human help, including software and security work.^[3] That same power, which thrilled developers, now looks dangerous to regulators who imagine foreign intelligence services steering the model to hunt for ways into American networks.^[1] The reported “jailbreak” was demonstrated by another company, raising alarms inside the Commerce Department that existing guardrails were not enough for such an advanced system.^[4] Instead of asking Anthropic to patch a narrow flaw, officials chose a sweeping response aimed at stopping any foreign use until they are satisfied the risk is under control.^[5]

Anthropic, for its part, stresses that it already built safety systems and logging into these models and calls the order a “misunderstanding” of both the evidence and the real risk.^[5] The company says it is “working to restore access as soon as possible,” which signals it does not believe Fable 5 and Mythos 5 are hopelessly unsafe but rather that the dispute is about thresholds and technical details. This clash fits a pattern in high-stakes technology: regulators act on early warning signs before outside experts can review the data, while companies argue that blanket bans overshoot the problem. Here, the Trump administration appears to be applying the same logic it has used on sensitive chips and cyber tools—better to restrict first and loosen later than wait for a foreign breach.^[2] For conservatives who favor strong national defense but limited government at home, this case raises a hard question: how to protect key tools from China, Russia, and others without punishing American users and innovators in the process.^[1]

Why Americans Lost Access Too — And What Comes Next

The most jarring twist for everyday users is that an order aimed at foreign nationals ended up cutting off access for everyone, including American citizens and companies.^[5] Anthropic explains that it simply cannot verify the nationality of every single user in real time across cloud providers and partner platforms.^[7] Instead of risking a violation of export law, it chose the blunt option: shut down Fable 5 and Mythos 5 to all customers worldwide, while leaving older models online.^[1] Amazon Web Services, which hosts many Anthropic systems, confirmed that Anthropic asked it to revoke access to these models “for all users in all regions,” underscoring how seriously the company takes the directive.^[1]

The US government just banned the world's most powerful AI model.

Not China. Not Russia. Britain. Canada. Australia. Everyone.

Yesterday at 5:21pm the Commerce Department sent Anthropic one letter.

By midnight Fable 5 and Mythos 5 were offline globally.

No warning. No… pic.twitter.com/zwvMrO05Ip

— Kuroyami (@kuroyamiai_) June 13, 2026

Social media posts and tech forums quickly filled with frustrated developers who had just started building tools or businesses on top of Fable 5 and Mythos 5, only to see them vanish overnight because Washington feared foreign abuse.^[6] Commentators also noted that foreign-born engineers working legally in the United States at Anthropic itself are now treated as “foreign persons” under the order, highlighting how broad these controls can reach.^[4] As legal challenges and talks continue, this episode sends a clear message: frontier artificial intelligence is now squarely in the middle of national security policy, and future crackdowns—right or wrong—will be felt not just in foreign capitals but on the phones and laptops of American citizens who rely on these tools for work, security, and daily life.^[3]