OpenAI hardware leader resigns over concerns about 'surveillance of Americans without judicial oversight and lethal autonomy without human authorization'

In the wake of OpenAI's debacle with the US Department of War, a lead of its robotics division has left the company, with a warning left on her X account for the AI company. In it, Caitlin Kalinowski says, "This wasn’t an easy call. AI has an important role in national security. But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got."

For context on that resignation, two weeks ago, Anthropic and the US Department of War had a falling out as the former refused to remove safeguards on its AI tools for use in weaponry and mass surveillance. In response, OpenAI stepped up to the plate to take the deal. Sam Altman did later clarify "we shouldn't have rushed to get this out on Friday", and that the Department of War wouldn't use OpenAI's tool in US intelligence agencies like the NSA.

At the time, Altman said: "There are many things the technology just isn’t ready for, and many areas we don’t yet understand the tradeoffs required for safety. We will work through these, slowly, with the DoW, with technical safeguards and other methods."

It seems that, from Kalinowski's post, it's not just about the potential of these tools to be used in the surveillance of Americans, but the fact that it could have even been on the table without adequate consideration. OpenAI started those talks, then Altman had to walk some of them back.

Kalinwoski says in a follow-up post that her problem was that the decision was rushed without defined guardrails. She says, "These are too important for deals or announcements to be rushed."

Despite this tension, Kalinowski does say, "This was about principle, not people. I have deep respect for Sam and the team, and I’m proud of what we built together."

As noted on Kalinowski's LinkedIn, she "led overall planning, hiring, and operating cadence for a fast-scaling robotics organization, built the systems that turned ambiguous work into executable programs, and partnered across the exec team, engineering, finance, real estate, IT, and operations to support physical-world scale-up."

As for what's next, Kalinwoski says that she is taking some personal time before moving on to something that is "building responsible physical AI". Still, one can hope that Kalinowski making her departure in such a public fashion with such strong words will push OpenAI to reflect on the deals it makes with the US government going forward.

It's worth noting that OpenAI has relied on government support in recent years. Trump's AI action plan is intended to cement US dominance in the AI field, which will naturally benefit OpenAI. The company has also argued that it should be allowed to scrape copyrighted content or risk losing out to Chinese AI, and it winning is in the interest of a very pro-AI US government. Though Altman said he is not looking for a government bailout if things go badly, continuing to sign deals with the US government would certainly help with the cash flow.