On March 7, 2026, Caitlin Kalinowski resigned as head of robotics at OpenAI, citing governance concerns related to the company’s new cooperation agreement with the U.S. Department of Defense.
The agreement, announced on February 28, allows the Pentagon to deploy OpenAI’s AI technologies within classified environments. OpenAI stated the partnership includes safeguards prohibiting domestic surveillance and autonomous lethal weapons.
Kalinowski publicly argued that the deal was finalized too quickly and lacked clearly defined guardrails. Her concerns focused on two risks: surveillance of U.S. citizens without judicial oversight and the possibility of lethal autonomy without human authorization.
She described the decision to resign as difficult, emphasizing that it reflected governance principles rather than personal conflict. Kalinowski also expressed respect for OpenAI CEO Sam Altman and pride in the robotics team’s work.
OpenAI has defended the defense partnership as a responsible approach to national-security applications of artificial intelligence.
Who Is Caitlin Kalinowski
Caitlin Kalinowski is a hardware engineering executive known for leading advanced consumer-technology programs. Before joining OpenAI, she spent several years at Meta Platforms, where she led development of augmented-reality glasses and other emerging hardware systems.
She joined OpenAI in November 2024 to establish and lead the company’s robotics initiative, bringing expertise in hardware systems, sensors, and product-scale device engineering. Her role focused on bridging large-scale AI models with physical hardware platforms.
OpenAI’s Robotics Initiative
OpenAI’s robotics effort is still in an early development stage but represents a strategic expansion from pure software into embodied AI.
The company operates a robotics lab in San Francisco that employs roughly 100 data-collection staff to train robotic systems through real-world interaction. The program focuses primarily on robotic arms performing household manipulation tasks.
The long-term objective is to enable general-purpose robotic capabilities powered by foundation models.
In December 2025, OpenAI announced plans for a second robotics lab in Richmond, California, intended to expand experimentation with hardware systems and large-scale data collection.
Under Kalinowski’s leadership, the robotics team emphasized practical engineering and real-world applications of AI in physical machines.
Industry Context
Kalinowski’s departure highlights ongoing tensions across the AI sector regarding cooperation between leading AI developers and military institutions.
As frontier AI systems increasingly influence defense capabilities, internal debates around governance, safeguards, and acceptable use cases are becoming more visible across major technology organizations.
Leave a comment