Senior technical staff from Anthropic are in Washington meeting with White House officials in an effort to resolve a dispute that led the company to take its most advanced artificial intelligence models offline, Axios reported Sunday, citing a person familiar with the matter.
The report could not be independently verified. Anthropic and the White House did not immediately respond to requests for comment.
According to Axios, Anthropic engineers have already held a series of virtual discussions with White House officials since initial outreach from the Trump administration on Friday, as both sides attempt to clarify the scope of new restrictions.
The dispute centers on a directive from the Trump administration requiring Anthropic to block access to its newest AI systems, Fable 5 and Mythos 5, for all foreign nationals, regardless of whether they are inside or outside the United States. In response, Anthropic said it disabled global access to the models.
The San Francisco-based AI company, which has confidentially filed for a U.S. initial public offering, had previously flagged security concerns about its Mythos model, citing its advanced cybersecurity and potential offensive hacking capabilities, and initially withheld it from broad public release.
Earlier this week, Anthropic released a public version of its Fable model, which the company said incorporated additional safeguards aimed at mitigating cybersecurity risks.
The ongoing dispute underscores growing tensions between artificial intelligence developers and U.S. regulators as governments move to impose tighter controls on advanced AI systems with national security implications.
