There Was Just an Unusually Unsettling Pentagon-Related Resignation at OpenAI

You probably won’t be surprised to learn that someone has resigned from OpenAI and has promoted his cause far and wide on social media. You probably hadn’t even heard of Caitlin Kalinowski before this moment, and her online statement, which reads in part, “I have deep respect for Sam and the team, and I’m proud of what we built together,” is not, in itself, interesting reading material.

Amid the company’s high-profile deal with the Pentagon, which CEO Sam Altman admitted “seemed opportunistic and reckless,” it should be no surprise that some employees have been laid off. OpenAI’s rival, Anthropic, appears to be on its way to victory, generating record revenues, after successfully promoting itself as an AI company that wants to save the world from mass surveillance and autonomous killer robots.

I mention all this because of Kalinowski’s specific role: She was the leader of the company’s robotics division.

I resigned from OpenAI. I care deeply about the robotics team and the work we create together. This was not an easy call. AI plays an important role in national security. But surveillance of Americans without judicial oversight and lethal autonomy without human authority…

– Caitlin Kalinowski (@kalinowski007) 7 March 2026

So when Kalinowski says in her statement that “lethal autonomy without human authorization” requires more deliberation than has been acknowledged, she speaks with at least some knowledge of what kind of AI-powered robots OpenAI has in mind.

According to a report by Wired last year, anonymous leakers revealed that OpenAI was looking to hire humanoid robot experts, or robots with at least partially human-inspired designs. It was also working towards creating AI algorithms that could “understand the physical world” and trying to “empower robots to navigate and perform tasks”.

Robotics-related job advertisements on the OpenAI website begin with the line “Our robotics team is focused on unlocking general-purpose robotics and moving toward AGI-level intelligence in dynamic, real-world settings.”

This doesn’t mean that any of OpenAI’s goals in robotics were accomplished, but when the person in charge of the most Terminator-like technology at your company leaves, and strongly implies that it happened because leadership was a little more comfortable with a guy who calls himself the Secretary of War, it can reasonably be considered a moment to do some soul-searching for those in the C-suite.