The U.S. government has ordered Anthropic to immediately shut down worldwide access to its two newest artificial intelligence models, Claude Fable 5 and Mythos 5, citing national security risks. The directive marks an unprecedented move by the Trump administration to restrict a homegrown American AI company's own technology from foreign users, including foreign employees working in the United States.

What Triggered the U.S. Government's Sudden Ban?

The export control order came less than a week after Anthropic released Claude Fable 5 and Mythos 5. According to reporting from Semafor, the government grew concerned after discovering that a China-linked group had potentially accessed Anthropic's new model. Additionally, the administration received a warning that Fable 5 could be "jailbroken," meaning hackers could bypass its safety guardrails to misuse the system.

David Sacks, an adviser to President Trump, revealed on social media that when Anthropic's co-founder and CEO Dario Amodei was notified about the jailbreak vulnerability, the company did not immediately fix it. However, Anthropic disputed this characterization, stating that the government provided only "verbal evidence of a potential narrow, non-universal jailbreak" and argued that such minor vulnerabilities exist in competing AI models as well.

The Pentagon's chief information officer, Kirsten Davies, defended the decision on social media, emphasizing that national security takes priority over corporate interests. "Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always," Davies stated.

Who Loses Access and Why Does It Matter?

The scope of the ban is remarkably broad. Foreign nationals cannot access Fable 5 or Mythos 5 whether they are inside or outside the United States. This includes foreign-born Anthropic employees themselves. Additionally, H-1B visa holders working in the U.S. and international research institutions collaborating with Anthropic will lose access to the company's most advanced models.

The practical consequences are significant. Global companies like S&P, an equity and research firm, use Anthropic's Claude software to integrate databases that help financial advisers and investors access data efficiently. International research institutions that partner with Anthropic will be cut off from cutting-edge AI technology. Even U.S.-based companies employing foreign workers will face productivity challenges when those employees cannot use the restricted models.

How This Reflects Broader U.S.-China AI Competition

The order underscores escalating tensions between the U.S. and China over artificial intelligence dominance. China's AI firm DeepSeek has made significant strides in recent years, releasing generative AI models DeepSeek-V3 and DeepSeek-R1 using a fraction of the cost compared to American counterparts. Beijing also controls rare earth supply chains, which are critical to the AI sector.

This move reflects the Trump administration's broader policy of export controls over high technology, particularly in critical sectors like AI and semiconductors. The U.S. government has previously imposed restrictions on chip sales by tech giants like NVIDIA and AMD. The Anthropic order signals that even American-developed software now faces strict gatekeeping based on who can access it.

Steps to Understand the Implications of AI Export Controls

• National Security Rationale: The U.S. government argues that advanced AI models could be weaponized or used for cyberattacks if they fall into hostile hands. Experts have told Reuters that Mythos models, in the wrong hands, could dramatically accelerate sophisticated cyberattacks, particularly in sectors like banking that rely on complex, decades-old technology systems.
• Enforcement Challenges: Tech community members have questioned whether the ban is practically enforceable. Using "foreign national" as the criteria to restrict access is difficult to enforce in practice and could be easily bypassed by people with genuine malicious intent, according to industry observers.
• Global Research Impact: The order prevents international collaboration on AI safety and development. Research institutions outside the U.S. that work with Anthropic will lose access to cutting-edge models, potentially slowing global AI progress and safety research.

Kun Chen, a former engineer at Meta, Microsoft, and Atlassian, commented on the policy's effectiveness. "Using 'foreign national' as the criteria to gate the model is just not very smart. It's clearly not enforceable in practice and would obviously lead to a wide ban," Chen stated.

What Does This Mean for Anthropic and the AI Industry?

Anthropic has said it believes there was a "misunderstanding" and is working to restore access to the models as soon as possible. The company had previously worked with the U.S. government on safety measures ahead of the Fable launch and noted that rival AI providers' models showed similar abilities to unearth minor bugs in code.

This order intensifies an already contentious relationship between Anthropic and the Trump administration. The San Francisco-based company is currently suing the administration after being placed on a supply chain blacklist for refusing to allow the U.S. military to use its AI models for domestic surveillance and fully autonomous weapons systems.

The move also highlights a paradox: while the U.S. government seeks to protect national security by restricting AI access, other major AI developers like OpenAI have generally avoided such strict direct restrictions from the Trump administration. Anthropic's situation appears uniquely contentious, driven by both security concerns and the company's refusal to cooperate with military applications.

Sridhar Vembu, co-founder of the Indian multinational company Zoho, observed that the order demonstrates how technology has become the ultimate geopolitical weapon. "National sovereignty, national security, all of it is now about technology," Vembu stated. He also emphasized that this situation presents an opportunity for countries like India to develop their own advanced AI capabilities.

As the U.S. tightens controls on its own AI technology, the broader question remains: will export restrictions accelerate innovation elsewhere, or will they help the U.S. maintain its technological edge? The answer may depend on whether Anthropic can resolve its disputes with the government and restore access to its newest models.