She Quit OpenAI on 'Principle' After Finding Out What They're Building for the Pentagon

Caitlin Kalinowski had one of the most coveted positions in artificial intelligence. She was helping build robots at OpenAI, the company behind ChatGPT. Then she quit.

Her reason? She says OpenAI moved too fast on a deal to give the U.S. military access to its AI systems, and nobody stopped to think hard enough about what that actually means.

In posts on social media, Kalinowski explained she resigned on "principle." Her biggest concerns were about two specific things: AI being used to spy on regular Americans without a court order, and AI-powered weapons that can decide to kill without a human pressing the button.

"AI has an important role in national security," she wrote. "But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got."

OpenAI pushed back, saying the Pentagon deal has clear "red lines" against domestic surveillance and autonomous weapons. But the fact that a senior team member felt those protections weren't strong enough to stay is raising serious questions.

This comes at a tense time. The Pentagon has been shopping around for AI tools from multiple companies. Anthropic, the maker of Claude, already clashed with Defense Secretary Pete Hegseth after its CEO refused to let the military use its AI for mass surveillance or autonomous killing.

The big picture: the people actually building these AI systems are starting to draw lines in the sand about how they should be used. And some of them are willing to walk away from dream jobs to make that point.

As reported by NPR.

Source: NPR