Trump Loosens Export of Anthropic AI Models, Mythos and Fable Open Again

Artificial intelligence (AI) illustration with the background of the United States flag. The US government lifted export restrictions on Anthropic's AI models Mythos and Fable to strengthen the competitiveness of American AI technology in the global market. Illustration made by AI.

JAKARTA - The United States government has lifted the export license requirement for two AI models owned by Anthropic, Mythos and Fable. This decision reopens access to the two models that were closed to the public after being included in the list of export-restricted technologies.

Launching a report by TechCrunch, Wednesday, July 1, Anthropic will begin restoring access to Mythos and Fable on Wednesday, July 1. Export licenses mean US government permits before certain technologies can be provided to foreign parties.

Previously, on June 12, the US government included the two models on a list of technologies that should not be provided to foreign citizens without special permission. The rule is difficult to implement on a large scale. As a result, Anthropic closed public access completely.

US Commerce Secretary Howard Lutnick said the easing was made after Anthropic agreed to a number of security commitments. The company will detect and address security risks, work with the US government to develop protocols and standards, and report on harmful activities related to the model.

However, according to TechCrunch, a number of cybersecurity experts have doubted the effectiveness of the restrictions from the start. Anthropic had previously promised to take many voluntary security measures, even before the export rules were published.

Some critics consider the ban to resemble a political pressure tool rather than a purely security measure. A number of Anthropic executives have previously criticized the way the government and President Donald Trump's political opponents could potentially use AI technology.

Mythos was first opened to a limited number of organizations in April. The model attracted attention for its ability to identify and exploit software security gaps. Another version called Fable was then released to the public in June with additional security fences.

Pressure on the US government increased after a number of Asian AI companies began releasing models with capabilities close to Mythos, including Fugu and Tulongfeng. Washington eventually loosened restrictions so that American-made AI models could remain competitive in the global market.

Last week, Lutnick also allowed Mythos to be released to certain customers approved by the White House. OpenAI's latest model was released to a group of organizations approved by the Trump team, not directly to the public.

The Trump administration's AI policy still leaves the industry operating in a gray area. AI companies have not received clear certainty about the rules that will apply to the launch of new models, especially after an executive order in June opened the door for the government to review models before they are released.