Defense Department and Technology Policy

Caitlin Kalinowski, a senior member of OpenAI’s robotics and hardware team, has resigned from the company, saying she left on principle after OpenAI moved ahead with a new agreement to deploy its AI systems inside secure U.S. Department of Defense computing environments. In public posts, she argued that policy guardrails were not adequately defined before the deal and said AI use for surveillance of Americans without judicial oversight and for lethal autonomy without human authorization are red lines that needed more deliberation. An OpenAI spokesperson told NPR the Pentagon agreement lays out a 'workable path' for national‑security uses of AI while making clear the company’s red lines of no domestic surveillance and no autonomous weapons, and acknowledged strong views among employees and the public. The resignation lands as federal agencies increasingly turn to OpenAI and Google for AI tools and after a very public clash between the Pentagon and rival firm Anthropic, whose CEO also rejected domestic mass surveillance and fully autonomous weapons uses for its models. The episode underscores a growing fault line between Washington’s demand for flexible AI tools in all 'lawful' operations and technologists who insist on hard limits around civil liberties and battlefield autonomy before their work is embedded in U.S. military systems.