Caitlin Kalinowski, who led hardware and robotics initiatives at OpenAI, has resigned following the company’s recent agreement with the U.S. Department of Defense to deploy its artificial intelligence models in classified environments.
Kalinowski announced her decision in a social media post, saying the company moved forward with the defense partnership without sufficient deliberation around governance and safeguards. The resignation comes amid broader debate over the role of AI companies in military and national security applications.
Resignation linked to Pentagon partnership
Kalinowski said the decision to leave OpenAI was difficult but driven by concerns about how the Pentagon agreement was handled.
“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.”
According to reports, the agreement allows OpenAI’s models to be used in classified Pentagon environments for national security-related work. The partnership was announced roughly a week before Kalinowski stepped down.
Despite her criticism of the decision-making process, Kalinowski expressed respect for OpenAI CEO Sam Altman and the company’s team.
OpenAI says contract includes safeguards
In response to concerns surrounding the agreement, OpenAI said its contract with the Department of Defense includes specific boundaries governing the use of its technology.
According to a company statement, the partnership establishes “red lines” that prohibit the use of OpenAI systems for domestic surveillance or fully autonomous weapons.
The company said the arrangement is intended to provide what it described as a responsible framework for applying artificial intelligence in national security contexts, while continuing discussions with employees, policymakers, and civil society groups.
Deal followed failed negotiations with Anthropic
The Pentagon partnership emerged after negotiations between the Department of Defense and AI startup Anthropic reportedly ended without an agreement over restrictions on how AI systems could be used.
Anthropic had sought guarantees that its technology would not be used for mass domestic surveillance or fully autonomous weapons. After those talks failed, the Pentagon reached an agreement with OpenAI instead.
OpenAI has said its own contract reflects similar safety principles, though critics have questioned whether the safeguards are sufficient or enforceable.
Internal tensions over AI and military use
Kalinowski joined OpenAI in 2024 to oversee robotics and consumer hardware initiatives after previously working on hardware development at Meta. Her departure highlights internal tensions around how AI companies engage with defense agencies.
Reports about the fallout surrounding the Pentagon deal suggest that some employees and industry observers have raised concerns about the potential for AI technologies to be used in surveillance or autonomous weapons systems, even if companies establish policy restrictions.
At the same time, OpenAI has argued that collaboration with government agencies can help ensure that AI systems are deployed responsibly in national security contexts.
Implications for OpenAI’s robotics work
Kalinowski’s resignation may also affect OpenAI’s developing hardware and robotics efforts. The company has been exploring physical AI systems, including robotics research and hardware prototypes, as part of its broader AI strategy.
While robotics is not the central focus of OpenAI’s current products, the company has been expanding its work in physical systems and related research areas.
The resignation underscores how decisions about military partnerships can ripple through AI organizations, influencing leadership, internal governance debates, and the future direction of emerging AI technologies.
Redactor
