Senior technical staff at Anthropic are preparing for a critical meeting at the Department of Commerce following an abrupt national security directive. The Trump administration ordered the AI firm to suspend foreign national access to its latest models, Fable 5 and Mythos 5, citing a method to bypass security safeguards.
The government alleges the jailbreak can identify software vulnerabilities. Anthropic counters that the flaws found are minor and detectable by other publicly available tools. In response, the company globally disabled access to the controversial models.
More than 80 cybersecurity executives from major firms signed an open letter to Commerce Secretary Howard Lutnick supporting Anthropic. The signatories urged the administration to lift the restrictions immediately.
The current rupture follows an earlier dispute where Anthropic refused to allow the U.S. military to use its AI for domestic surveillance and fully autonomous weapons. The San Francisco startup, which has confidentially filed for an IPO, had previously delayed releasing the Mythos model over its potential hacking capabilities.