Sam Altman from OpenAI recently expressed a desire to help mediate the ongoing conflict between the Pentagon and its primary AI competitor, Anthropic. The tension escalated when Anthropic announced it would not ease military security protocols before a critical deadline this Friday.
In a memo to staff late Thursday, Altman noted that OpenAI shares similar safety “red lines” with Anthropic, emphasizing that their AI technologies should not facilitate mass surveillance of civilians or operate weapons autonomously.
He also mentioned ongoing discussions with the Department of Defense about a potential agreement that could allow the deployment of OpenAI’s models in sensitive military applications. Altman highlighted that these talks are crucial for establishing a responsible framework and preventing a risky precedent within the industry.
“This has evolved into an industry-wide matter, and it’s essential that we clarify our stance,” Altman wrote, reiterating the importance of collective responsibility. He indicated that OpenAI could reach an understanding with the Pentagon on modeling usage that aligns with their ethical guidelines.
“This debate centers around control rather than the application of AI. While corporations have significant influence, we believe that no private entity should wield more power than a democratically elected government. Democracy is complex, but we remain steadfast in our commitment to it,” he added.
Earlier reports highlight that Anthropic’s Claude chatbot, the sole AI currently being utilized within classified military operations, is facing a deactivation deadline of 5 p.m. ET on Friday. Pete Hegseth has threatened to designate Anthropic as a “supply chain risk” or invoke the Defense Production Act to mandate compliance unless they adhere to the Pentagon’s stipulations.
This memo followed a statement by Anthropic CEO Dario Amodei, who remarked that the company couldn’t responsibly meet the Pentagon’s demands. He further critiqued the latest proposal from the Pentagon, suggesting it failed to address concerns regarding mass surveillance or the potential use of Claude in fully autonomous weapon systems.
Emil Michael, the Undersecretary of Defense for Research and Engineering, responded sharply to Amodei’s position, claiming he has a “God complex” and prioritizes personal control over national security.
At this intersection of rising tensions, the relationship between Altman and Amodei appears strained. Notably, the two recently avoided a handshake during a photo opportunity at an AI summit in India. Meanwhile, it has been reported that Elon Musk’s Grok has received approval for classified military use, with OpenAI and Google being on the brink of similar arrangements.
