Developer says OpenAI silently kicked him out of cyber Trusted Access program after full identity verification

QuixiAI · x · 2026-09-05

A developer accepted into OpenAI's Trusted Access for Cyber (TAC) program is airing a messy ordeal: after verifying his identity via Persona (license, facial biometrics, passport) and being told to buy physical hardware by September 1 to stay in the program, he was quietly removed. OpenAI first called it an "error," then told him to reapply — after which Persona, on OpenAI's behalf, said it could not verify him. Support stated it "cannot manually override a Daybreak verification decision." He notes the contradiction: "a 'decision' is not an 'error'" — OpenAI says it corrected a technical issue in the Daybreak Blue verification process that removed access for some accounts, who must now reapply. The episode highlights how little human recourse exists in OpenAI's high-sensitivity access programs.

Original post →

More from Models

Models channel →