A top executive at OpenAI has left the company over its agreement with the Defense Department that allows its technology to be deployed in the military.
Caitlin Kalinowski, the employee who led OpenAI’s hardware and robotics efforts, announced her resignation on social media on Saturday.
“This was not an easy call. AI has a critical role in national security,” Kalinowski wrote do. “But surveillance of Americans without judicial oversight and lethal autonomy without human authority are lines that require more deliberation than has been discussed.”
“It was about principle, not about people,” he said. He stressed that he still has “deep respect” for CEO Sam Altman.
Last month, OpenAI announced a new deal with the Pentagon, with rival Anthropic facing threats from top Trump officials for not coming to a similar deal. Negotiations between Anthropic and the Pentagon had failed because, according to CEO Dario Amodei, the company insisted on mass surveillance of US citizens and a ban on its AI systems being used in autonomous weapons without humans in the loop. When it became clear that Anthropic would not back down, the Pentagon shut down the company and followed through on its threat to declare it a “supply chain risk”, which prevents it from signing any military contracts.
This turned into a public relations disaster for OpenAI, with many criticizing Altman for messing with an extremely unpopular and aggressive administration. Originally, OpenAI had agreed that its AI systems could be used for “all lawful purposes”, and only after heated feedback did Altman say it would do so. Update Pentagon Deal To include specific protections on surveillance and autonomous weapons.
In short, OpenAI looked like it was selling out while Anthropic remained like a hero. This news triggered a mass exodus of users from OpenAI’s ChatGPT to Anthropic’s cloud, with the cloud app usurping ChatGPT from the top of the App Store.
Disagreements also emerged within the industry. More than 1,000 former and current employees of OpenAI and Google have signed an open letter demanding their employers reject the Pentagon’s demands to use AI technology for mass surveillance and autonomous weaponry.
Kalinowski explained his departure, saying that his “issue is that the announcement was rushed without being defined.”
“This is first and foremost a governance concern,” he explained. a follow up post. “These are too important to rush into deals or announcements.”
Altman himself admitted that the deal was “rushed” and that its “optics don’t look good.”
in a statement Provided to media outletsOpenAI emphasized its updated agreement with the Pentagon and its clear red lines: “No domestic surveillance and no autonomous weapons.”
“We recognize that people have strong views about these issues and we will continue to discuss them with employees, government, civil society and communities around the world,” the company said.
Kalinowski’s departure and the broader uproar over OpenAI’s military agreement come as concerns over the use of AI in warfare, and especially in decisions that could claim human lives, are more urgent than ever. Despite Anthropic’s public resistance to DOD, reports suggest that the cloud was used to select targets for deadly US missile strikes on Iran. Although not confirmed, it remains the frightening possibility that one of those identified targets was a girls’ school in Minab, where at least 168 people, the majority of them children, were horrifically killed. Killed by an airstrike that was probably a US Tomahawk missile.
More on OpenAI: Anger over OpenAI has increased so much that entire protests are taking place against it.
