On June 12 local time, American AI company Anthropic announced that the U.S. government has issued instructions prohibiting any foreign entities from accessing two of its AI models. This move has caused dissatisfaction among some people in the U.S. AI industry, who complain that President Donald Trump has allowed the sale of certain AI chips to China, but is unwilling to allow American allies to access advanced AI models.
According to the US newspaper The New York Times, Anthropic company stated in a statement that the U.S. government has issued export control orders targeting the Mythos 5 and Fable 5 models on the grounds of ‘national security’. A source familiar with the matter revealed that these directives came from the U.S. Department of Commerce, and it is not yet clear how long the restrictions will last.
The statement states that Anthropic received instructions from the U.S. government at 17:21 local time on the 12th, but the U.S. government did not provide details regarding the “national security issues” in the letter. According to the order, Anthropic must revoke access rights to Mythos 5 and Fable 5 models for all foreign citizens, including its foreign employees, to ensure compliance.
Anthropic stated in its statement that it would comply with the government’s lawful directives and has revoked all user access to both AI models. However, the company publicly questioned the United States government’s decision, opposing the restriction of a commercial model on the grounds of “potential jailbreak vulnerabilities”.
On June 9th, local time, Anthropic released the Fable 5 and Mythos 5 models. Both models are based on the same underlying architecture. Fable 5 is the first Mythos-level model that is open to the public, while Mythos 5 retains more complete capabilities and is still only accessible to a few carefully selected institutions. The company boasts that Mythos has "excellent performance" in the field of cybersecurity, capable of efficiently identifying network vulnerabilities.
June 9th, Anthropic announced the release of the Claude Fable 5 and Claude Mythos 5 models.
The New York Times reported that Anthropic provided the Mythos model to approximately 40 organizations that maintain critical computer infrastructure for patching security vulnerabilities. Researchers at these organizations have varying reactions: some view Mythos as a concerning new threat, while others believe it is merely a routine technological upgrade.
The order issued by the U.S. government has caused dissatisfaction among some people in the AI industry in the United States. Dean Ball, a researcher at the American Innovation Foundation who served as an AI advisor during the Trump administration, complained that this decision is "unintelligible".
Bauer said, "The government allows the sale of certain models of AI chips to China on one hand, while prohibits countries like the UK from using our best models on the other hand. I don't know what to say."
Bauer said that he cannot determine whether this is an escalation of the legal dispute between the U.S. government and Anthropic, or some form of “extreme national security stance.” However, such behavior is extremely absurd and could affect the deployment of advanced AI models in the United States. He criticized on social media, saying, “If the advanced AI industry doesn’t exist, there’s no need to worry about advanced AI models being abused.”
Anthropic and the US Department of Defense are facing a legal dispute. During negotiations for a contract worth approximately $200 million, there was disagreement over how AI tools should be used in warfare. Anthropic demanded that the Pentagon ensure that its AI models would not be used for large-scale surveillance within the United States, nor to power completely autonomous lethal weapons. The Pentagon argued that private companies have no right to formulate policies on behalf of the U.S. government.
After the negotiations collapsed, U.S. Defense Secretary Hagel announced that Anthropic was classified as a "supply chain risk" company. On March 9 local time, Anthropic filed a lawsuit against the U.S. Department of Defense, accusing it of misusing the "supply chain risk" label to improperly impose punitive measures on U.S. companies.
The New York Times reports that the Trump administration has been indifferent towards AI, but after Anthropic released its Mythos model, the White House began considering introducing new regulatory measures. On June 2 local time, Trump signed an executive order requiring AI companies to provide government with access rights for up to 30 days before the new model is officially released. This directive is not mandatory, and companies can choose to participate voluntarily.
