OpenAI revokes access for some cybersecurity researchers in TAC program after technical glitch

Key Points

• OpenAI temporarily revoked TAC access for a limited number of vetted cybersecurity researchers.
• The issue affected the Daybreak Blue tier, which provides advanced AI models for defensive security tasks.
• OpenAI cited a technical error on its side and asked affected users to re‑verify and reapply.
• All impacted researchers reported residing outside the U.S. and Europe, hinting at possible regional constraints.
• The incident underscores ongoing tension between AI guardrails and the needs of legitimate security research.