The U.S. Commerce Department has ordered artificial intelligence company Anthropic to suspend global access to its latest AI models, Mythos and Fable 5, citing concerns that the technology could be exploited by military intelligence agencies in countries such as China and Russia. Commerce Secretary Howard Lutnick informed Anthropic CEO Dario Amodei in a letter that the government feared the models could be used in ways that threaten U.S. national security. Following the directive, Anthropic disabled access to the models worldwide and entered negotiations with federal officials.
The dispute centers on allegations that Fable 5’s cybersecurity safeguards can be bypassed through a “jailbreak” technique, potentially allowing users to identify software vulnerabilities. Anthropic acknowledged that a bypass method exists but argued that it exposes only minor flaws and that similar capabilities are available through other public AI models. The company stated it had worked closely with the government to test Fable 5 before its release and had received approval to deploy the model publicly earlier this month.
The Commerce Department invoked powers under the 2018 Export Control Reform Act, marking the first known use of those authorities to restrict an AI model. The move has sparked debate among export control experts, who question whether AI systems accessed remotely can legally be classified as exports. More than 80 cybersecurity leaders, including executives from Nvidia and Adobe, have signed an open letter urging the Trump administration to lift the restrictions, while discussions between Anthropic and U.S. officials continue.
Pic Courtesy: google/ images are subject to copyright
