OpenAI robotics chief resigns over considerations about Pentagon AI deal : NPR

[ad_1] <img src="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/5448x3632+0+0/resize/1100/quality/50/format/jpeg/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Fc4%2F06%2Fff796a644b44af98293238717310%2Fgettyimages-2225604584.jpg" data-template="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/5448x3632+0+0/resize/{width}/quality/{quality}/format/{format}/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Fc4%2F06%2Fff796a644b44af98293238717310%2Fgettyimages-2225604584.jpg" class="img" alt="OpenAI CEO Sam Altman speaks in Washington, D.C., on July 22, 2025." fetchpriority="high"/> OpenAI CEO Sam Altman speaks in Washington, D.C., on July 22, 2025. Mandel Ngan/AFP through Getty Photographs disguise caption

toggle caption

Mandel Ngan/AFP through Getty Photographs A senior member of OpenAI's robotics group has resigned, citing considerations about how the corporate moved ahead with a lately introduced partnership with the U.S. Division of Protection.

Caitlin Kalinowski, who served as a member of technical workers targeted on robotics and {hardware}, posted on social media that she had stepped down on "precept" after the corporate revealed plans to make its AI programs out there inside safe Protection Division computing programs.

The settlement is a part of a broader push by the U.S. authorities to include superior AI instruments into nationwide safety work, a development that has sparked debate throughout the tech business about oversight and acceptable makes use of.

In public posts explaining her resolution, Kalinowski wrote: "I resigned from OpenAI. I care deeply concerning the Robotics group and the work we constructed collectively. This wasn't a straightforward name."

She mentioned coverage guardrails round sure AI makes use of weren't sufficiently outlined earlier than OpenAI introduced an settlement with the Pentagon. "AI has an essential position in nationwide safety," Kalinowski wrote. "However surveillance of People with out judicial oversight and deadly autonomy with out human authorization are traces that deserved extra deliberation than they received."

Kalinowski additionally emphasised that her considerations had been extra concerning the course of quite than particular executives inside the corporate, saying she had "deep respect for Sam and the group, and I am happy with what we constructed collectively," referring to OpenAI chief government Sam Altman.

A spokesperson for OpenAI advised NPR the corporate believes the settlement with the Pentagon "creates a workable path for accountable nationwide safety makes use of of AI whereas making clear our crimson traces: no home surveillance and no autonomous weapons."

It continued, saying the corporate acknowledged "individuals have robust views about these points and we'll proceed to interact in dialogue with staff, authorities, civil society and communities around the globe."

Kalinowski's resignation comes amid heightened competitors amongst main builders of synthetic intelligence to produce expertise to the U.S. authorities. In latest weeks, federal companies have turned to OpenAI and Google for AI programs as tensions elevated with a rival agency, Anthropic, over the army use of its fashions.

Anthropic's CEO spoke out towards permitting the corporate's software program for use for functions similar to home mass surveillance or autonomous weapons, a stance that led to clashes with protection officers — together with Secretary of Protection Pete Hegseth — who mentioned the division wants flexibility to deploy business AI instruments in all "lawful" operations.

Inside OpenAI, Kalinowski's position targeted on constructing out the corporate's robotics group because it scaled. She wrote on her LinkedIn profile that this included hiring to assist the corporate's enlargement into AI efforts tied to bodily infrastructure and equipment.

Kalinowski signaled she plans to proceed working in the identical area. "I am taking a bit of time, however I stay very targeted on constructing accountable bodily AI," she wrote.