OpenAI locked researchers out of TAC program after technical error
OpenAI accidentally locked some researchers out of its Trusted Access for Cyber (TAC) program because of a technical error.
TAC is meant for vetted users who use AI tools to spot security threats, but this week, several people got surprise notifications saying their accounts weren't eligible.
OpenAI confirmed the issue and asked those affected to reapply.
Five Daybreak Blue researchers affected
All five of those locked out were researchers outside the US and Europe using the new Daybreak Blue tier, which lets them try advanced models like GPT-5.6 Sol for finding vulnerabilities and analyzing malware.
OpenAI said it was an issue on our end, and not the user experience we want to deliver, and that it was due to a technical issue affecting a limited number of users.
The company says it's still committed to supporting cybersecurity research, just with re-verification to maintain their access.