The OpenAI and Anthropic Execs at the Center of the Pentagon Action

OpenAI was on the backfoot on Monday. Current and former OpenAI employees took to social media to share concerns about their employer signing an agreement with the Pentagon that seemed to accept at least some terms Anthropic had rejected. A surge in people uninstalling the ChatGPT app from their phones over the weekend, according to TechCrunch, suggested consumers were also responding. 

The ChatGPT creator responded quickly. On Monday night, OpenAI announced it had updated its Pentagon contract to specify that its models may not be used to conduct domestic surveillance via commercially obtained data. And OpenAI CEO Sam Altman said in a post on X that the Pentagon affirmed that OpenAI’s services “will not be used by Department of War intelligence agencies (for example, the NSA).” 

Inside OpenAI, Altman, national security partnerships chief Katrina Mulligan, and technical staff member Boaz Barak, who has focused on AI safety, tried to calm employee concerns by explaining the company’s Pentagon deal in greater detail, we are told.

What comes next depends on a handful of people at these companies and the Department of Defense: