AI Safety Director Submits Identical Memo For Third Consecutive Year, Updates Only The Date

SAN FRANCISCO—Rhiannon Alcott, who joined Apex AI in 2023 as Director of Responsible AI Deployment, confirmed Wednesday that she has submitted her annual risk memo for the third consecutive year with no changes beyond the date at the top.

"I updated 2025 to 2026," said Alcott, 34, who holds a PhD in cognitive science and spent two years at an AI alignment non-profit before accepting a role the company described at the time as "pivotal." "The executive summary is identical. The five recommendations are identical. Last year one of them got a thumbs-up in Slack."

Alcott's position, which carries the title Director and reports to a Senior Vice President who reports to a Chief Risk Officer who reports to the General Counsel, does not include budget authority, team membership, product-review veto power, or access to model weights. Her quarterly presentation slot was moved to an async Loom recording in Q2 of last year following a scheduling conflict that has not been resolved.

She noted that three of the five scenarios she flagged in her 2023 memo — AI-generated legal filings containing hallucinated case law, open-weight models enabling influence operations, and deployment timelines outpacing internal safety evaluations — have since become front-page news.

The company's most recent press release described its responsible AI approach as "industry-leading" and listed Alcott's department under a section titled "Ongoing Commitment."

"I'm grateful for the platform," Alcott said. "I have an office. It has a window. The window faces a parking garage."