OpenAI Robotics Lead Resigns Over Pentagon Deal, Citing Surveillance and Lethal Autonomy Concerns (2026)

Incident Summary

Caitlin Kalinowski, OpenAI’s Head of Robotics and Consumer Hardware, resigned on March 7, 2026, approximately one week after OpenAI announced a deal to deploy its AI models on the Pentagon’s classified computing network.^[1] The deal followed the Pentagon’s decision to ban Anthropic from federal contracts after Anthropic refused to remove safeguards against domestic surveillance and autonomous weapons use.^[3]

In her resignation statement posted to X, Kalinowski wrote: “I resigned from OpenAI. I care deeply about the Robotics team and the work we built together. This wasn’t an easy call. AI has an important role in national security. But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got.”^[4]

In a follow-up post, she clarified: “To be clear, my issue is that the announcement was rushed without the guardrails defined. It’s a governance concern first and foremost. These are too important for deals or announcements to be rushed.”^[1]

Kalinowski had joined OpenAI in November 2024. Previously she led AR glasses hardware at Meta (the Orion project) and spent approximately six years at Apple designing MacBooks.^[1]

The resignation was the most prominent individual departure triggered by the Pentagon deal, distinct from the broader #QuitGPT consumer boycott movement that saw approximately 2.5 million participants and a 295% surge in ChatGPT uninstalls (see INC-26-0034). An open letter signed by more than 900 employees from OpenAI and Google demanded that their employers reject all Pentagon surveillance contracts.^[2]

OpenAI stated it maintained “red lines” against domestic surveillance and autonomous weapons. Sam Altman reportedly acknowledged the contract was “opportunistic and sloppy” and renegotiated its terms.^[3]

Key Facts

• Governance concern, not anti-military stance: Kalinowski explicitly stated “AI has an important role in national security” and framed her objection as a process failure: “the announcement was rushed without the guardrails defined”^[4]
• Two specific lines cited: Surveillance of Americans without judicial oversight, and lethal autonomy without human authorization^[4]
• 900+ employee open letter: Employees from both OpenAI and Google signed an open letter demanding rejection of Pentagon surveillance contracts^[2]
• Altman acknowledgment: CEO reportedly described the original contract as “opportunistic and sloppy”^[3]
• Prior role: Head of Robotics and Consumer Hardware, with senior hardware leadership experience at Meta and Apple^[1]
• Context: Pentagon had just banned Anthropic for refusing to remove safety restrictions, creating commercial opportunity that OpenAI moved to fill^[3]

Threat Patterns Involved

Primary: Safety Governance Override — Kalinowski’s resignation statement describes a governance process failure: a consequential military AI deployment was announced before guardrails were defined. The “governance concern first and foremost” framing indicates that internal review processes for safety-critical decisions were either bypassed or insufficient.

Secondary: Accumulative Risk & Trust Erosion — This resignation adds to a pattern of senior personnel departing AI companies over safety governance concerns in early 2026, including departures from Anthropic (Mrinank Sharma), OpenAI (Mission Alignment Team dissolution), and the broader #QuitGPT movement.

Secondary: Lethal Autonomous Weapon Systems — Kalinowski specifically cited “lethal autonomy without human authorization” as a line that warranted more deliberation. The Pentagon deployment of frontier AI models raises questions about the boundary between decision support and autonomous action in military contexts.

Significance

1. Insider governance critique: Unlike external critics, Kalinowski had direct visibility into OpenAI’s internal decision-making process. Her characterization of the deal as rushed and lacking defined guardrails carries evidentiary weight about the state of internal governance at the time of the announcement.
2. Speed vs. governance tension: The incident illustrates the tension between commercial opportunity (filling the gap created by Anthropic’s exclusion) and governance process (ensuring safety-critical deployments have defined guardrails). Altman’s reported acknowledgment that the contract was “opportunistic and sloppy” validates this framing.
3. Distinct from opposition to military AI: By explicitly affirming that “AI has an important role in national security,” Kalinowski distinguished her objection from a blanket anti-military stance. Her concern was procedural: governance structures should precede deployment announcements for military AI, not follow them.
4. Employee voice as safety signal: The resignation, combined with the 900+ employee open letter, indicates that internal concern about governance gaps existed broadly, not as an isolated dissenting view. Senior departures function as a safety signal that external observers can use to assess organizational governance health.