A small group of vetted experts suddenly faced identity checks and fresh applications, raising questions about how regional access controls work.
Several cybersecurity professionals have complained about the sudden revocation of access to OpenAI’s Trusted Access for Cyber (TAC) program. It allows vetted researchers to use artificial intelligence models with fewer restrictions for cybersecurity-related tasks.
OpenAI confirmed that the access issues were caused by a technical error. On the Cyber page in ChatGPT, users saw messages stating that their identity could not be verified or that their accounts did not meet the program’s requirements.
What the TAC Program Provides
Trusted Access for Cyber is intended for professionals engaged in legitimate security research. Participants can use OpenAI’s latest models to identify vulnerabilities, review software code, analyze malware, and report discovered issues to developers so they can release fixes.
Anthropic has a similar mechanism for vetted researchers: the Cyber Verification Program. Such programs are intended to distinguish defensive cybersecurity work from potentially dangerous uses of AI tools.
Access to Daybreak Blue Had to Be Reapplied For
One researcher received an email saying that their access to Daybreak Blue – the new TAC tier for individual professionals – had been revoked. OpenAI asked them to submit a new application and complete the identity verification process again.
This was an issue on our side, not the user experience we aim to provide.
– OpenAI email provided by the researcher
Five professionals affected by the access revocations live outside the United States and Europe. This may indicate that the issue affected users in certain regions.
OpenAI explained that a small group of users no longer had active access to Daybreak Blue. To restore or retain it, they need to complete verification again.
Daybreak Blue and Daybreak Red Tiers
OpenAI introduced Daybreak Blue on August 10 as an entry-level tier for most defensive cybersecurity professionals. It provides access to general-purpose models, including GPT‑5.6 Sol, with safeguards for permitted defensive tasks.
At the same time, the company launched the higher Daybreak Red tier. It is designed for authorized vulnerability research, exploit validation, and security testing.
The TAC access incident occurred amid criticism of OpenAI and Anthropic from cybersecurity researchers. They emphasize that overly strict restrictions in AI models can sometimes hinder legitimate work needed to identify and eliminate digital threats.
