The Trump administration has imposed export controls on Anthropic's most advanced artificial intelligence models, barring foreign governments, companies, and individuals from accessing the technology. Commerce Secretary Howard Lutnick notified Anthropic CEO Dario Amodei on Friday that the Mythos 5 and Fable 5 models would face restrictions on any transfer outside the United States or to foreign nationals within the country.
The escalation reflects Washington's hardening stance on treating state-of-the-art AI systems as critical national security infrastructure. It also deepens Anthropic's complicated relationship with federal regulators, who have already designated the company's work as too sensitive for some government applications while simultaneously demanding control over its commercial deployment.
The Commerce Department's decision came after another company reportedly discovered a method to jailbreak the Mythos model, triggering alarms about potential security vulnerabilities. Administration officials said they initially asked Anthropic to delay releasing the latest versions, but the company proceeded with the launch anyway. That prompted the export control letter as a forceful backup measure.
Under the new regime, Anthropic must obtain licenses for any export, re-export, or domestic transfer of the affected models. The company must also submit applications for individually validated licenses before proceeding with transfers. Violations carry financial and civil penalties.
An administration official indicated the restrictions are intended as temporary leverage. The models will remain locked down until the U.S. government's national security infrastructure reaches a higher readiness threshold, a process the official suggested could take several weeks.
The move sits uneasily with the Trump administration's broader AI policy. Earlier this month, the White House issued an executive order mandating testing of advanced AI systems before deployment, but the order is voluntary and deliberately avoids a licensing framework. David Sacks, the White House chief AI adviser, engineered that approach to prevent what he views as regulatory overreach that could stifle innovation.
An administration official stated that President Trump wants to avoid harming the AI industry and intends for innovation to continue unimpeded. The export controls on Anthropic's models suggest that commitment has limits when national security appears at stake.
Anthropic did not provide comment on the Commerce Department action.
Author James Rodriguez: "The irony is sharp: Washington claims to want a thriving AI sector while simultaneously locking down the most promising models before their commercial value can fully materialize."
Comments