Summary: Dario Amodei, chief executive of Anthropic, is scheduled to hold a meeting with White House chief of staff Susie Wiles on Friday, a move that has been reported as signaling progress in the artificial intelligence firm's dispute with the Pentagon. The discussions come as U.S. officials acknowledge technical strengths in Anthropic's latest model, Mythos, particularly for defensive cybersecurity tasks.
According to reporting, the possible meeting follows acknowledgements within the administration of Mythos' sophisticated capabilities to assist in cyber defense and breach response. The models and programs under discussion were described as having utility specifically for defensive cybersecurity purposes.
White House and Anthropic spokespeople did not immediately reply to requests for comment. The report that first flagged the scheduled meeting has not been independently verified by Reuters.
Sources close to the negotiations were quoted as saying it would be "grossly irresponsible" for the U.S. government to forgo the technological advantages offered by the new model, adding that such a course could advantage China. That characterization was reported by Axios and attributed to a person involved in the talks.
Anthropic introduced Mythos on April 7 and has been rolling the technology out under "Project Glasswing," described as a tightly controlled program that allows selected organizations to use an unreleased Claude Mythos Preview model strictly for defensive cybersecurity purposes.
Bloomberg reported that the U.S. government plans to make a version of Mythos available to major federal agencies. Separately, Anthropic co-founder Jack Clark stated on Monday that the company was in discussions about Mythos with the Trump administration, even after the Pentagon severed business ties with the AI lab amid a contract dispute.
This sequence of developments - the planned White House meeting, the administration's recognition of Mythos' defensive capabilities, and reports of controlled deployments to select organizations and federal agencies - frames an unfolding dialogue about how advanced AI systems might be integrated into national cybersecurity defenses while negotiations continue with the Pentagon.
Contextual note: The available reporting identifies the actors, their public positions, and the programs involved but does not provide independent confirmation of the meeting or the detailed terms of any agreements.