Emerging
Jun 17, 20261
60%
Trump Administration Orders Anthropic to Block AI Model Access, Raising Questions About Unclear Export Rules
The Trump administration ordered Anthropic to block access to its newest AI models, Fable 5 and Mythos 5, citing national security concerns related to a potential vulnerability exploitable by groups linked to China. The order marks the first known use of US export controls to restrict access to an AI model, raising questions among experts about whether existing regulatory frameworks adequately address cloud-based AI services and creating uncertainty across the frontier AI industry.
Quick Facts
Who
Anthropic
What
Trump administration issued export control directive against Anthropic
When
June 2026 (week of June 17)
Where
United States
- Trump administration issued export control directive against Anthropic
- Anthropic blocked access to Fable 5 and Mythos 5 models
- Government cited national security authorities as justification
- Government alleged jailbreak vulnerability exploitable by China-linked groups
- Anthropic disputed the severity of the vulnerability
The Trump administration has ordered Anthropic to restrict access to its newest AI models, Fable 5 and Mythos 5, forcing the company to block access for all foreign nationals and even its own international employees. The government cited "national security authorities" to justify the "export control directive," though the administration has not publicly explained the legal basis for the order. Anthropic stated that the government's concerns involved a potential "jailbreak" that could be exploited by groups linked to China, though the company disputed whether this vulnerability actually circumvented its safeguards.
Experts describe the order as unprecedented and highlight fundamental uncertainties in how export controls apply to cloud-based AI services. Traditionally, export controls have targeted tangible goods—weapons, hardware, and discrete digital items like software or source code that can be physically transferred. The Anthropic case presents a different scenario: the models remain hosted on Anthropic's servers, and users access them remotely through a chatbot interface rather than receiving model weights, source code, or executable copies. Hanna Dohmen, a senior research analyst at Georgetown University's Center for Security and Emerging Technology, told reporters, "To my knowledge, this is the first time US export controls have been used to control access to an AI model in this way."
The lack of clarity has left AI developers uncertain about their obligations. Andrew Reddie, a professor at UC Berkeley's Goldman School of Public Policy, noted that while export control rules give the government "wide latitude" to restrict access to certain goods, "the equivocation by successive administrations regarding what the responsibilities of model developers are" has created confusion across the industry. The order raises immediate questions for other frontier AI labs, including OpenAI, Google, Meta, and xAI, about what restrictions they may face as they develop more capable models.
Policy experts warn that relying on opaque, ad hoc interventions to govern AI development is unsustainable. The ambiguity surrounding remote access to cloud-based AI services represents a known gap in current export control regimes—one that Congress is attempting to address through legislation under consideration in the Senate. Until clearer rules are established, companies operating in the AI sector face significant regulatory uncertainty.
Why This Matters
This order establishes a precedent for government intervention in frontier AI development and reveals critical gaps in existing export control frameworks. Companies developing advanced AI models—including OpenAI, Google, Meta, and xAI—now face significant regulatory uncertainty about what international access restrictions they may be forced to implement. The lack of transparent legal criteria creates operational chaos: developers cannot plan compliance strategies, and policymakers lack clarity on how to govern rapidly evolving AI capabilities. Without legislative clarification, ad hoc government directives will likely proliferate, making it harder for US companies to operate competitively while potentially undermining legitimate international collaboration.
Timeline & Sources
Jun 17, 2026
WireTrump administration issues export control directive against Anthropic's Fable 5 and Mythos 5 models