OpenAI Executive Resigns Following Pentagon Partnership Announcement
One of OpenAI’s senior executives leading the development of robotics hardware, Caitlin Kalinowski, has resigned from the company.
The decision came after OpenAI announced a partnership with the Department of Defense concerning the use of artificial intelligence models.
Through a post on the X platform (formerly Twitter), Kalinowski stated that the partnership announcement was conducted too hastily without clear policy guardrails.
“My concern is that the announcement was rushed without clear safeguards. This is a fundamental governance issue. These things are too important to be done hastily in a deal or announcement,” said Kalinowski.
“This is about principles, not about people. I deeply respect Sam (OpenAI CEO) and the team, and I am proud of what we have built together,” Kalinowski added.
OpenAI confirmed Kalinowski’s resignation. In a statement to media outlet Engadget, the company said they have no plans to appoint a replacement for the former head of robotics hardware in the near term.
The company also explained that they understand some parties hold strong views regarding AI use in military and national security contexts. According to OpenAI, their partnership with the Pentagon includes clear limitations.
“We believe our agreement with the Pentagon creates an appropriate pathway for responsible AI use for national security purposes, whilst clarifying our boundaries: no domestic surveillance and no autonomous weapons,” OpenAI stated.
Kalinowski’s resignation is believed to be the most notable consequence of OpenAI’s decision to enter into a partnership with the US Department of Defense.
This is because the OpenAI-Pentagon partnership had previously faced fierce criticism from social media users, who felt OpenAI was supporting military operations, particularly concerning mass surveillance and the use of AI for autonomous weapons.
For context, OpenAI signed the Pentagon partnership agreement in February. The partnership concerns the use of OpenAI’s AI in “secure systems” or “classified networks” at the Pentagon.
OpenAI’s decision also contrasts with its competitor Anthropic, which has actually refused to loosen certain restrictions on AI use, according to reports from Engadget.