OpenAI's robotics division has recently experienced leadership changes. Caitlin Kalinowski, who oversees hardware in its robotics division, announced her resignation just as the company finalized a controversial AI agreement with the U.S. Department of Defense. Kalinowski joined OpenAI from Meta in 2024. On social media platform X, she stated that while AI plays an important role in national security, certain boundaries require more cautious public discourse. "Surveillance of Americans without judicial oversight, and enabling lethal autonomy without human authorization—these are issues that deserve more careful consideration ," she wrote. Kalinowski emphasized that her departure was based on principled grounds, expressed continued respect for OpenAI CEO Sam Altman, and voiced pride in the team's work in robotics. OpenAI later confirmed her resignation and defended its collaboration with the Pentagon. A company spokesperson explained that the agreement aims to establish pathways for the responsible application of AI in national security while clearly defining red lines: no domestic surveillance and no development of autonomous weapons. The spokesperson also noted that the company will continue discussions with employees, government bodies, and society at large regarding these matters. This partnership comes after competitor Anthropic declined similar agreements. Previously, Anthropic opted out due to concerns that its technology could be used for mass surveillance or autonomous weaponry. In response, U.S. President Donald Trump criticized the company via Truth Social and directed federal agencies to cease using its technologies. Additionally, U.S. Secretary of Defense Pete Hegseth designated Anthropic as a supply chain risk, restricting defense contractors from collaborating with it. The deal between OpenAI and the Pentagon sparked backlash among users as well. Some indicated they would stop using ChatGPT, while Anthropic's chatbot, Claude, rose to first place on the free apps chart of the Apple App Store last month. Reports indicate that downloads of Claude in the United States increased by approximately 240 percent month over month in February. Kalinowski's exit introduces further uncertainty into OpenAI's robotics initiatives. Over the past year, the company quietly established a laboratory in San Francisco employing around 100 data collectors to train robotic arms for tasks such as household chores. These projects were seen as part of OpenAI's exploration of humanoid robots and hardware directions; however, according to informed sources, current hardware plans remain in early stages and have not become core business priorities for the company.