OpenAI's Robotics Hardware Lead Resigns Over Pentagon Deal, Citing Rushed Guardrails on Surveillance and Autonomous Weapons
Caitlin Kalinowski left OpenAI after the company signed a classified Pentagon agreement without sufficient deliberation on safeguards against domestic surveillance and lethal autonomy.
Overview
Caitlin Kalinowski, the hardware executive who led OpenAI’s robotics team, resigned on March 7 in protest of the company’s agreement to deploy its AI models on a classified Department of Defense network, according to NPR. Kalinowski said that safeguards against domestic surveillance and autonomous weapons were not adequately defined before the deal was announced, calling her departure a matter of principle.
The resignation marks the most prominent internal departure over OpenAI’s accelerating push into defense work, and comes amid an ongoing industry reckoning over the boundaries of AI in military applications.
What We Know
Kalinowski joined OpenAI in November 2024 after leading the team that built Meta’s Orion augmented reality glasses, as reported by Fortune. At OpenAI, she oversaw hardware and robotics operations, hiring staff and scaling the company’s physical AI ambitions.
She announced her departure on X, writing: “I resigned from OpenAI. I care deeply about the Robotics team and the work we built together. This wasn’t an easy call,” according to NPR.
Her core objection centered on process rather than the principle of defense collaboration itself. “AI has an important role in national security,” she wrote, as quoted by NPR. “But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got.”
She further described the rollout as “rushed without the guardrails defined,” characterizing it as “a governance concern first and foremost,” according to Engadget. She emphasized that her objection was “about principle, not people,” and that she maintained deep respect for CEO Sam Altman and her colleagues.
OpenAI confirmed the departure and told NPR that the Pentagon agreement “creates a workable path for responsible national security uses of AI” while establishing “clear red lines: no domestic surveillance and no autonomous weapons.” According to Engadget, the company said it has no plans to replace Kalinowski in the role.
Context: The Pentagon Deal and Anthropic’s Refusal
The resignation sits within a broader chain of events that has divided the AI industry. OpenAI signed the Pentagon agreement on February 27, just hours after the Department of Defense designated rival Anthropic a supply-chain risk over its refusal to allow its models to be used for domestic mass surveillance or fully autonomous weapons, as reported by Fortune.
Anthropic’s negotiations with the Pentagon had collapsed after CEO Dario Amodei pushed for strict limitations on those same two use cases. The rapid sequence of events—Anthropic’s blacklisting followed by OpenAI’s deal—drew criticism that OpenAI had capitalized on its competitor’s principled stance.
CEO Sam Altman later acknowledged that the announcement appeared “opportunistic” and said he would amend the deal to explicitly prohibit spying on Americans, according to Engadget.
The deal also triggered a consumer backlash: ChatGPT uninstalls surged 295 percent in the week following the announcement, and rival Claude briefly topped the Apple App Store before Anthropic’s servers buckled under the demand.
What We Don’t Know
Several questions remain unresolved. The specific terms of OpenAI’s classified Pentagon agreement have not been made public, making it difficult to evaluate the company’s stated red lines against actual contractual language. It is unclear whether Altman’s pledge to amend the deal has resulted in binding changes or remains a public commitment.
The long-term impact on OpenAI’s robotics program is also uncertain. With no plans to replace Kalinowski, it is unclear who will lead the hardware and robotics team or whether the company’s physical AI ambitions will be scaled back.
Whether Kalinowski’s departure will prompt further resignations within OpenAI or across the broader industry remains to be seen. Her exit is the most senior principled departure since the Pentagon deal was announced, but rank-and-file sentiment inside the company has not been publicly reported.
Analysis
Kalinowski’s resignation underscores a tension that has been building across the AI industry for months: the gap between companies’ stated ethical commitments and the speed at which defense contracts are being signed. Her framing of the issue as a governance failure—not an objection to national security work per se—highlights a distinction that may become increasingly important as more AI firms enter the defense market.
The episode also illustrates how the Anthropic-Pentagon standoff has created a competitive dynamic in which ethical positioning and commercial opportunity are deeply intertwined. OpenAI’s rapid move to fill the void left by Anthropic’s refusal has secured a major government contract, but at the cost of internal credibility and public trust among a segment of its user base that valued the company’s original safety-first identity.