Thought Leadership
Something unprecedented has happened: OpenAI rival Anthropic has been forced to suddenly restrict access to its most advanced AI model after the U.S. government raised national security concerns. Many details remain unclear.
Following an order from the U.S. government, AI company Anthropic has blocked access to its newly launched flagship AI software just days after its release. Government agencies instructed Anthropic, citing national security concerns, to prevent all foreign nationals from accessing the company’s AI models Fable 5 and Mythos 5, according to the company.
The restriction also applies to foreign nationals who are currently in the United States, including Anthropic employees. As a result of the directive, the company said it had to temporarily cut off access for all affected users. According to Anthropic, the government believes it has identified a method that could potentially bypass restrictions built into the software.
High-Profile AI Capabilities
The AI technology behind Anthropic’s Mythos model is particularly effective at detecting software vulnerabilities that may have remained undiscovered for years or even decades. Until now, the technology had been used by U.S. government agencies and selected companies to identify and patch security flaws.
However, one concern from the beginning has been that such an AI system could become a dangerous cyber weapon if it fell into the wrong hands.
The Fable 5 model, released only last week, is based on Mythos technology. However, its cybersecurity and biotechnology capabilities are disabled. Mythos 5 is the non-public full version, which is still intended to be used only by government agencies and selected enterprise partners to strengthen their systems.
Unclear Threat Assessment
Anthropic emphasized that it has only received limited information from the government so far. The company reviewed a report that it believes was likely the reason behind the order. According to Anthropic’s experts, the issue involved a limited capability that allows the AI to analyze specific pieces of code and fix errors.
Similar capabilities are also available in models from other providers, including OpenAI’s rival model GPT-5.5. Anthropic said it disagrees with blocking software used by hundreds of millions of users based on this concern. The company stressed that Fable 5’s security safeguards had been extensively tested.
Anthropic Faces Growing Tensions With U.S. Government
The situation is particularly notable because Anthropic CEO Dario Amodei had only days earlier argued that the government should be able to block potentially dangerous AI software. However, Anthropic insists that such decisions must be based on transparent processes, clear criteria, and technical evidence. The company argues that those conditions were not met in this case.
Anthropic has repeatedly made headlines in recent months. The company risked a conflict with the U.S. government after refusing to allow its AI models to be used in autonomous weapons systems or for mass surveillance in the United States, despite political pressure.
The Pentagon subsequently classified Anthropic as a supply chain risk, a move that could significantly restrict the use of the company’s software across government agencies. Anthropic is challenging the decision in court. At the same time, Mythos’ exceptional cybersecurity capabilities have once again made the company strategically important to the U.S. government.
dpa
