Lab's Head of Responsible AI Has Recommended Against 12 Deployments; All 12 Shipped

SAN FRANCISCO—Three years into her role as Head of Responsible AI at a frontier lab she is contractually prohibited from naming in interviews, Dr. Mara Stoll has signed off on 847 model deployments, attended 1,246 cross-functional alignment meetings, and authored four safety reports that have appeared in Senate testimony, three regulatory filings, and a Forbes profile of her CEO headlined "The Man Building the Future Responsibly."

She has recommended against 12 deployments. All 12 shipped.

"I have a seat at the table," she said at an AI governance conference last week, a phrase that appears in 23 of her public appearances since 2023. She paused. "It's a very large table."

Her team is herself and a contractor who formats the safety reports for external publication. The lab's capabilities division employs 340 full-time researchers, occupies three floors of the same building, and maintains a Slack channel, #responsible-ai, with 912 members and, per an internal audit she commissioned herself, four who have posted in the past 90 days.

In Q2, she flagged a consumer-facing deployment for potential harm to adolescent users. It was approved in three hours. The announcement called it "a watershed moment for human-centered AI."

"I wrote that phrase," she said.

Asked whether she planned to remain in the role, she said she was committed to "driving real change from within the institution"—a line that also appears in her 2023 offer letter, her 2024 performance review, and a note taped to her monitor she described as "motivational, mostly."

She is the most cited name in all four of the lab's congressional submissions. The lab announced last quarter it is hiring aggressively in safety and alignment. Of 34 open roles, 33 are in capabilities.

The contractor handles formatting for those job postings, too.