Story Highlights
- Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei directing the company to suspend all foreign-national access to its Fable 5 and Mythos 5 AI models.
- Anthropic disabled both models for all users worldwide, including domestic customers, stating it had no technical way to selectively restrict access.
- The action follows a reported warning from Amazon that researchers had successfully jailbroken the Mythos-class model to extract sensitive cybersecurity information.
What Happened
Anthropic disabled access to its most advanced artificial intelligence models, including Mythos, following an unprecedented order by the Trump administration to keep the technology out of the hands of all foreign nationals. The U.S. government told Anthropic to suspend access to the Fable 5 and Mythos 5 models by any foreign national “whether inside or outside the United States,” citing national security concerns.
Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei saying that the Mythos 5 and Fable 5 models would be subject to export controls to any location outside of the U.S. and to all foreign persons within the country. An administration official told Axios the Commerce Department decided to take the action after another company claimed it was able to jailbreak Mythos, alarming the administration about possible national security risks.
According to multiple media reports, Amazon chief executive Andy Jassy first raised concerns about the model with senior administration officials on Thursday after Amazon researchers used a series of prompts to get the Mythos-class model to provide information about cyberattacks that was supposed to be restricted.
Anthropic rolled out its latest AI model, Claude Fable 5, earlier this week, claiming it represents a new level of capability it calls “Mythos-class” — the first of that tier made available to ordinary users. The company claimed at launch that the model’s capabilities exceed those of any model it had ever made generally available. The model was particularly effective at identifying software vulnerabilities.
Anthropic complied with the directive but disputes it, arguing the standard would halt all new frontier model deployments across the AI industry. The company believes the government acted after learning of a jailbreak technique that exploits relatively minor vulnerabilities other models can also discover.
Why It Matters
This action is without precedent in the history of American technology regulation. The U.S. government has previously used export controls to limit the sale of semiconductor chips, restrict software exports, and block foreign investment in sensitive technology sectors. But ordering the immediate shutdown of an already-deployed commercial AI model — one being actively used by paying customers around the world — is a categorically different exercise of government power.
The order marks one of the administration’s most aggressive steps yet to control access to frontier AI models and significantly increases tensions with Anthropic, which has become a prominent voice in Washington policy circles for its public commitments to AI safety. The move appears to stem from concerns about a possible jailbreak of Anthropic’s systems.
The action also fits into a broader pattern of conflict between the Trump administration and Anthropic specifically. President Trump in February ordered all federal agencies to stop using Anthropic’s models after the company refused to agree to the Pentagon’s preferred contract terms for AI vendors, which stipulated that any AI models it purchased could be used “for any lawful purpose.” Anthropic had been seeking exemptions from having its models used for autonomous weapons systems or mass domestic surveillance. The Pentagon declared Anthropic a “supply chain risk” in early March.
For American businesses and developers who rely on these models for commercial operations, the shutdown is a stark reminder that AI infrastructure now sits within the reach of executive national security authority. What was treated as a commercial product one day can be restricted by government order the next, with little advance warning and no clear appeal process.
Economic and Global Context
The move marks the first time the U.S. government has used export controls to halt access to a commercial AI model already widely used. That distinction matters enormously for the technology industry. Export controls have long applied to hardware and proprietary code, but applying them to a live, deployed AI service creates a new category of government intervention in the digital economy.
Per Commerce’s letter, a license will be required for the export, re-export, or domestic transfer of those Anthropic models. The administration tried to get Anthropic to pause releasing the latest models but was unsuccessful, which prompted the export control letter. The model needs to remain locked down until the U.S. government’s national security apparatus is hardened, with officials suggesting that could happen in the next few weeks.
Anthropic’s competitive position in the global AI market has been significantly damaged by the episode. International customers and enterprise clients in Europe and Asia who had integrated these models into their workflows are now frozen out. Rival AI developers — including those based in China — have faced no comparable restrictions, creating an uneven playing field that critics say ultimately disadvantages American companies.
Implications
The administration’s action sets a precedent that will define AI governance for years. Every major AI company now understands that its frontier models are subject to executive national security review, and that the government can force a global shutdown based on a threat it may not fully disclose. That dynamic will influence how companies design, deploy, and communicate about their most capable systems going forward.
One official described the new arrangement as “a de facto licensing regime,” adding that companies “will not screw with the White House.” That framing suggests the administration views this not merely as a one-time national security intervention, but as the establishment of a new operating norm for frontier AI deployment in the United States.
For Anthropic, the path forward runs directly through Washington. Senior technical staff traveled to D.C. over the weekend to meet with White House officials, signaling the company’s intent to negotiate a path back to full deployment. Whether that path involves accepting government review requirements, agreeing to pre-deployment testing protocols, or something else entirely will shape the competitive landscape for the entire AI industry.
For American liberty and constitutional governance, this episode raises questions that have not yet been fully litigated: what due process rights, if any, do companies have when facing executive orders that eliminate their commercial product overnight? As AI becomes more deeply embedded in American economic and civic life, those questions will become increasingly urgent.
Sources
“Scoop: Trump admin blocks foreign access to Anthropic’s most powerful AI”
