In mid-June 2026, the U.S. Commerce Department issued a letter that forced Anthropic to immediately shut down access to Claude Fable 5 and Mythos 5 for every user on the planet, citing national security concerns. The company had no practical way to verify each user's nationality in real time, so it disabled the models entirely within hours. This unprecedented move raises fundamental questions about whether the U.S. government has the legal authority to control global access to artificial intelligence systems, and whether it exercised that power responsibly.

What Exactly Happened Between Anthropic and the U.S. Government?

On June 12, 2026, the Commerce Department's Bureau of Industry and Security sent Anthropic a letter ordering the company to prevent all foreign nationals from accessing Claude Fable 5 and Mythos 5, Anthropic's two most capable large language models (LLMs), or AI systems trained on vast amounts of text data to understand and generate human language. The order cited national security concerns and gave Anthropic essentially no time to comply. Because the company had no way to instantly verify the nationality of every user accessing the models from around the world, Anthropic made the only practical choice: it shut the models off for everyone.

The shutdown lasted until June 30, 2026, when the Commerce Department agreed to lift its restrictions after Anthropic significantly strengthened the models' safety guardrails, or built-in restrictions designed to prevent harmful outputs. However, this came at a steep cost. The new safety controls were so restrictive that they caused what observers described as a "collapse" of the models' benchmark scores, meaning the AI systems performed noticeably worse on standard tests of intelligence and capability. In one experiment, the updated Fable 5 completed only 3 out of 12 tasks that would have been routine before the new restrictions were imposed.

Did the U.S. Government Act Legally?

The Commerce Department justified its action by citing the Export Control Reform Act of 2018, a law originally written to regulate the export of physical hardware like uranium enrichment centrifuges. This was the first time the government had used export controls to prevent foreign access to an AI service, raising serious legal questions about whether the law even applies to AI models at all.

A legal scholar specializing in technology law and policy identified two major unresolved legal issues with the government's action:

• Whether AI access qualifies as an export: When a user sends a prompt to Fable 5 or Mythos 5 and receives a response, the only thing that leaves Anthropic's servers is the text reply. The model itself never leaves the company's infrastructure. The Commerce Department's own past guidance has treated remote access to software running on U.S. servers as outside the scope of export controls, suggesting the 2018 law may not apply to AI model outputs at all.
• Whether proper procedures were followed: The Commerce Department used an "is informed" mechanism, which is normally reserved for notifying a company that transactions to a specific country require government approval. Applying this to block all foreign nationals everywhere on the planet may exceed the department's legal authority under existing law.
• Lack of traditional safeguards: The order was unilateral, secret, open-ended, and global in scope. By contrast, the government's traditional approach to controlling dangerous technologies uses a temporary classification that requires sign-off from the Defense and State departments, is published publicly, expires after one year unless renewed, and commits the government to reviewing restrictions with U.S. allies.

What is clear is that the legal foundation for the government's action remains uncertain. Congress is currently considering legislation to clarify and expand the government's power to restrict AI access, which itself suggests that existing law may not provide the authority the Commerce Department claimed to exercise.

Why Didn't Anthropic Fight Back in Court?

Notably, Anthropic did not challenge the order's legality. Instead, the company complied immediately and later called the episode "a misunderstanding." Company officials then traveled to Washington not to litigate but to negotiate the restriction. This decision may have been strategic rather than a sign of corporate fear. Anthropic had previously sued the Trump administration over a supply chain risk designation, so the company clearly has a willingness to pursue legal action when it chooses.

One key reason Anthropic may have opted for negotiation over litigation is that the company largely agrees with the principle of government oversight of advanced AI systems. Just two days before the Commerce Department's letter arrived, Anthropic CEO Dario Amodei published an essay arguing that the government "should have the power to block or deter deployment" of frontier AI models, or cutting-edge systems considered too dangerous to release. Going to court to deny the government that power would directly contradict Anthropic's own public position.

Additionally, the 2018 statute strips federal courts of their usual power to overturn such decisions as arbitrary or unreasonable. Anyone challenging a directive in court must instead prove that the order was flatly unauthorized or unconstitutional, a far narrower and more difficult legal path.

What Does This Mean for the Future of AI Regulation?

The Anthropic case marks a significant shift in how the U.S. government attempts to control access to potentially dangerous new technologies. Traditionally, the government's regulatory process is deliberately slow and transparent. Multiple agencies consider possible controls, request public comment, coordinate restrictions with U.S. allies, and publish the resulting regulations in the Federal Register. The Anthropic order bypassed nearly all of these safeguards.

The coming legal and policy fight, should it arise, will not be over whether the government can exert this level of control; it already has demonstrated that it can. Instead, the fight will center on how governments can wield this control responsibly and transparently. The best-case scenario is that this fight takes place in the open, with public input, through lawful legislative processes. The dystopian alternative is a two-tiered AI order in which governments condition export privileges on secret access to frontier models more powerful than anything publicly known or available.

For now, the incident has sent a chill through the U.S. AI industry, signaling that even the most advanced AI companies can face sudden, sweeping restrictions with minimal legal process or advance notice.

" }