Caitlin Kalinowski, the director of hardware in OpenAI's robotics team, has left her position. In a post on X, she expressed frustration with the organization's swift agreement to work with the Department of Defense, pointing out the lack of thorough safeguards in place. OpenAI informed Engadget that it has no intention of filling her vacancy.
Before moving to OpenAI toward the end of 2024 from a role at Meta, Kalinowski shared on X her concerns that practices like monitoring U.S. citizens absent court supervision and self-directed lethal systems without operator approval warranted far greater scrutiny. In a reply to another user, the ex-executive noted that the partnership reveal proceeded too quickly before establishing necessary protections, framing it primarily as a matter of internal oversight.
OpenAI acknowledged her departure and, in comments to Engadget, recognized that individuals hold firm opinions on such matters, pledging to maintain dialogues with stakeholders. The firm also clarified in its response that it opposes the specific problems highlighted by Kalinowski.
According to OpenAI's official position, the arrangement with the Pentagon outlines a framework for ethical applications of AI in national defense, explicitly drawing boundaries against internal monitoring and independent weaponry.
Kalinowski's exit stands as a prominent consequence of OpenAI's pact with the Department of Defense, which followed Anthropic's rejection of requests to ease restrictions on widespread monitoring tools and self-operating armaments. OpenAI's chief executive, Sam Altman, has indicated willingness to revise the defense agreement to bar any intelligence gathering on U.S. residents.