Caitlin Kalinowski resigns from OpenAI as concerns rise over its Pentagon deal, emphasizing the need for scrutiny on AI's role in national security.
Caitlin Kalinowski has stepped down from her role as head of hardware at OpenAI following the company's contentious agreement with the U.S. Department of Defense. Her resignation highlights growing concerns about the implications of AI in national security.
“This was not an easy decision. Artificial intelligence plays a crucial role in national security. However, surveillance of Americans without a warrant and lethal autonomy without human approval are issues that deserve more scrutiny than they are currently receiving,” Kalinowski stated.
Before joining OpenAI in 2024, Kalinowski held positions at Apple and Meta, where she was instrumental in the Reality Labs division. She clarified that her resignation was principled rather than personal, expressing admiration for OpenAI CEO Sam Altman and the team.
“In my opinion, the announcement was made hastily, without defining boundaries. These are primarily governance issues. They are too important to rush into agreements,” she added.
OpenAI defended its deal with the Pentagon, asserting it aims to establish responsible AI use in national security while setting clear limitations against internal surveillance and autonomous weapons. An OpenAI representative acknowledged the strong feelings surrounding these topics, committing to ongoing dialogue with employees and the public.
The agreement was finalized in late February, shortly after negotiations with Anthropic faltered. Altman later recognized that the deal was rushed and could negatively impact OpenAI's reputation.
This controversy has affected user engagement with OpenAI's ChatGPT, which saw a 295% increase in deletions, while Claude, a competitor, gained popularity, topping App Store charts.
Anthropic's Challenges with the Pentagon
The U.S. Department of Defense had previously awarded contracts worth up to $200 million to several companies, including Anthropic, for AI security solutions. However, reports emerged that disagreements over Anthropic's strict ethical policies could lead to contract termination. The company prohibits the use of its Claude model for surveillance and autonomous military operations.
In February 2026, the U.S. Army reportedly utilized Claude in a mission related to Venezuelan President Nicolás Maduro, escalating tensions. Anthropic's CEO, Dario Amodei, expressed a preference not to collaborate with the Pentagon under conditions that could compromise democratic values.
Following President Donald Trump's directive to cease using Anthropic technologies, the Pentagon labeled the company a "supply chain risk," affecting its business relationships. Despite this, Microsoft and Google pledged to continue supporting Anthropic.
Legal Actions Ahead
Amodei announced plans to legally challenge the Pentagon's classification of Anthropic as a supply chain risk, arguing it lacks legal grounding. He emphasized that the majority of Claude's clients remain unaffected by this decision.
Following a leaked internal memo that criticized OpenAI's dealings with the Defense Department, Amodei apologized, indicating it was not intended for public disclosure. He described the memo as an outdated assessment and reiterated the company's commitment to providing essential tools to national security personnel.
“It is not in our interest to exacerbate the situation,” he stated.
While Anthropic plans to continue supplying its models to the Defense Department at cost, experts note the challenges of contesting government decisions related to national security.
