A high-ranking hardware executive at OpenAI has stepped down due to the company’s recent pact with the US Department of Defense, expressing concerns that the agreement lacked adequate precautions.
Caitlin Kalinowski, who oversaw OpenAI’s robotics and consumer hardware division, announced her resignation through social media posts on Saturday. She had joined OpenAI in 2024 after leading the development of augmented reality hardware at Meta.
Kalinowski highlighted the significance of AI in national security but raised alarms about the potential surveillance of Americans without proper legal oversight and the deployment of lethal autonomous systems without human approval. She stressed that her decision was driven by principles and not personal relationships, commending CEO Sam Altman and the team. In a subsequent post, she clarified that her primary issue was the rushed nature of the announcement, lacking defined safeguards. She underscored the importance of governance in such matters, asserting that critical decisions should not be hastily made.
OpenAI acknowledged Kalinowski’s departure to TechCrunch, affirming their belief that the collaboration with the Pentagon establishes a viable framework for responsible national security applications of AI, with clear boundaries against domestic surveillance and autonomous weaponry. The company recognized the contentious nature of these issues and committed to ongoing dialogues with employees, government entities, and civil society.
The partnership between OpenAI and the Pentagon was made public just over a week ago after discussions between the Pentagon and competing AI company Anthropic collapsed. Anthropic had been seeking assurances against widespread domestic surveillance and the development of fully autonomous weapons. Consequently, the Pentagon categorized Anthropic as a risk to the supply chain.