The Trump administration has reportedly softened its stance on Anthropic's powerful Mythos AI model, which was previously subjected to an outright ban on use by foreign nationals. This development follows a period of heightened tension between the AI company and the government, stemming from national security concerns over Mythos's advanced cybersecurity capabilities. While the full scope of this partial lifting of the ban is not immediately clear, it is understood that specific exemptions are being made for organizations and individuals directly involved in cybersecurity defense and national security initiatives.
This move comes after Anthropic was compelled to suspend all access to its Mythos 5 and Fable 5 models globally, just three days after their general release, due to an export control directive citing national security. The government had previously expressed concerns about the model's ability to exploit cybersecurity vulnerabilities, a capability Anthropic itself had highlighted when initially limiting Mythos's release. Despite the previous ban and Anthropic's blacklisting as a "supply chain risk" earlier in the year, the White House has maintained communication with the company, indicating the strategic importance of its AI technology.
The initial ban was triggered by what some Trump administration officials described as Anthropic's "recklessness" and insufficient responsiveness to government concerns regarding the models' potential misuse. Specifically, reports indicated that Amazon AI experts were able to "jailbreak" the Fable 5 model, unlocking the full cyber capabilities of the more advanced Mythos during initial testing. The current partial lifting suggests a potential compromise or a recognition of Mythos's utility in specific, controlled environments for national defense.
