Several security researchers reported that OpenAI revoked their access to the Trusted Access for Cyber (TAC) program, which provides vetted cybersecurity professionals with access to advanced AI models for research. The issue was attributed to a technical error, according to OpenAI, which confirmed the problem stemmed from an internal mistake. Researchers from outside the U.S. and Europe said they encountered the issue when attempting to access the TAC program through the ChatGPT Cyber page, where they received messages indicating their accounts were ineligible or their identities could not be verified. OpenAI asked affected users to reapply and complete the verification process. The company also cited a recent technical issue in a tweet, stating that a limited set of users’ access to Daybreak Blue, the latest tier of TAC, is no longer active. The revocations have raised concerns among researchers, who live outside the U.S. and Europe, suggesting the issue may be region-specific. OpenAI launched Daybreak Blue on August 10, offering access to frontier general-purpose models like GPT-5.6 Sol with safeguards tailored for defensive security work. The company also introduced a higher-tier program called Daybreak Red for more advanced cybersecurity research. Researchers have previously criticized the guardrails imposed by OpenAI and Anthropic, arguing that they hinder legitimate work. OpenAI referred TechCrunch to a tweet explaining the issue.
OpenAI’s TAC program allows vetted researchers to access advanced AI models with fewer cybersecurity guardrails than standard users, enabling them to report bugs and vulnerabilities to companies. The goal is to help defenders identify and patch flaws more quickly while preventing malicious actors from exploiting these models. The TAC program is similar to Anthropic’s Cyber Verification Program (CVP), which also grants access to advanced models for cybersecurity research. Both programs aim to balance the need for security with the ability of researchers to conduct legitimate work. The revocation of access to Daybreak Blue has raised questions about the stability and reliability of these programs, especially for researchers who rely on them for critical cybersecurity tasks. OpenAI’s response to the issue has been to ask affected users to reapply and complete the verification process, which has left some researchers uncertain about the future of their access.
The researchers who reported the issue said they live outside the U.S. and Europe, indicating that the revocations may be limited to certain regions. OpenAI’s tweet stated that a limited set of users’ access to Daybreak Blue is no longer active, and they will need to re-verify to maintain their access. The company did not provide further details about how many users were affected or the exact cause of the issue. The revocation has sparked discussions among cybersecurity professionals about the challenges of balancing security and research access in AI programs. OpenAI’s response to the issue has been to ask affected users to reapply and complete the verification process, which has left some researchers uncertain about the future of their access.
Source: techcrunch
