Anthropic cuts foreign access to Fable 5 and Mythos 5 after US order
Anthropic said US agencies told it to block all foreign nationals from two advanced AI models over national security concerns.
By Daniel Okafor · Business Editor
3 min read
Anthropic said it has shut off foreign nationals’ access to two advanced AI models after receiving an order from US government agencies. The move matters because the company says the directive applies even to foreign nationals who are in the United States, including some of its own staff.
In a Friday blog post, Anthropic said officials instructed the company to bar all foreign nationals from using Fable 5 and Mythos 5, citing national security concerns. Anthropic said it received the order at 5:21pm, or 21:21 GMT, on Friday.
The company said the government letter did not give a detailed explanation of the security issue behind the order. Anthropic said it had to act quickly and cut off access on short notice.
Al Jazeera and Reuters reported that the restriction covers foreigners currently in the US as well as users abroad. Anthropic said that group includes people working at the company.
Cybersecurity concerns around Mythos
Anthropic’s Mythos AI system can identify software vulnerabilities, including flaws that have gone undiscovered for long periods, according to Al Jazeera and Reuters. US authorities and selected companies have used that capability to close security gaps, they reported.
The same capability has raised concerns that advanced AI tools could be misused as cyberweapons, according to Al Jazeera and Reuters. Anthropic said it has so far received only partial information from the government about the basis for the order.
Fable 5, which Anthropic released this week, is built on Mythos technology, according to the company. Anthropic said Fable 5 does not allow access to its cybersecurity and biotechnology capabilities.
Mythos 5 is the full non-public version of the system, according to Al Jazeera and Reuters. They reported that it is meant to remain available only to government agencies and selected corporate partners for work aimed at strengthening systems.
Company challenges the basis for the order
Anthropic said it reviewed a report that it believes may have led to the government action. The company said its experts concluded the report appeared to describe a limited ability to use AI to examine specific program code and fix errors.
Anthropic said other AI systems, including OpenAI’s GPT-5.5, also have that capability. The company said it disagreed with blocking software used by hundreds of millions of people on that basis.
Anthropic also said Fable 5’s safety measures had undergone extensive testing. The company did not say when it expected the restriction to be reviewed or lifted.
Earlier in June, Anthropic called for leading AI companies to coordinate a pause in the development of advanced systems, warning that rapid gains in the technology could create a risk that humans lose control. In that blog post, the company said the world should have the option to slow or temporarily pause development as AI systems become faster at carrying out tasks.
This story draws on original reporting from Al Jazeera.