Several security researchers reported on OpenAI’s support forums and on X that their access to the Trusted Access for Cyber (TAC) program was abruptly revoked. When they opened the Cyber page in ChatGPT they saw a message stating that their identity could not be verified or that their account was 'ineligible at this time.' OpenAI later confirmed the revocation resulted from an internal error affecting a limited number of users.
TAC gives vetted cybersecurity defenders access to OpenAI’s capable models while applying fewer safety restrictions than the offering. Entry requires submitting identification and passing a vetting review; Anthropic runs a comparable effort called the Cyber Verification Program (CVP). The stated purpose of both initiatives is to let trusted parties find and report software flaws faster, narrowing the window for malicious abuse. One researcher said OpenAI emailed them that their Daybreak Blue tier access was withdrawn 'due to a technical issue affecting a limited number of users,' noting the mistake was on OpenAI’s side.
- TAC provides reduced‑guardrail access to OpenAI models for security researchers.
- Participation requires ID submission and a vetting process; Anthropic’s CVP works similarly.
- The recent revocation was traced to a technical error, not a policy change.
Why this matters
Although the source cites an internal error affecting a limited user set, the event shows how vendor‑managed access programs can create single points of failure for security work. Delays in vulnerability reporting and eroded trust may follow, indicating a need for clearer communication, backup access routes, and independent eligibility checks.