Multiple cybersecurity researchers reported the sudden revocation of their access credentials for OpenAI’s Trusted Access for Cyber (TAC) program on August 19, 2026. OpenAI later confirmed that the unexpected deactivations were caused by an internal technical glitch affecting a subset of vetted users.
Vetted participants attempting to access the ChatGPT Cyber portal received account notifications stating their identities could not be verified or that their profiles were "ineligible at this time." In communication shared with affected researchers, OpenAI stated that access to Daybreak Blue—the defensive tier of TAC—was removed due to an internal system error, prompting users to re-complete identity verification.
The issue predominantly affected vetted researchers based outside of the United States and Europe.
The Structure of Frontier Cyber Tiers
OpenAI established the Trusted Access for Cyber initiative alongside frontier model rollouts to resolve tensions between dual-use AI risks and legitimate vulnerability defense. Standard model guardrails frequently refuse queries involving network scanning, reverse engineering, and exploit proof-of-concepts, which can impede defensive security teams.
The program operates across two primary access tiers launched earlier in August:
- Daybreak Blue: Designed for defensive cybersecurity practitioners, incident responders, and software auditors. It grants access to frontier architectures, including GPT-5.6 Sol, with relaxed refusal boundaries tailored for malware analysis, patch verification, code auditing, and vulnerability triage.
- Daybreak Red: A higher-clearance tier providing dedicated offensive security models for vetted red teams conducting authorized penetration testing and exploit validation.
Similar clearance frameworks have been deployed by peers across the industry, such as Anthropic’s Cyber Verification Program (CVP). Both programs require researchers to submit government-issued identification and organizational credentials to undergo manual and automated vetting.
Balancing Guardrails with Defensive Research
The technical disruption comes amid ongoing industry debates surrounding the latency and false-positive rates of frontier AI safety classifiers. Cybersecurity professionals have frequently reported that automated refusal filters block benign defensive analysis, while frontier lab red-teaming reports continue to highlight the risk of autonomous exploit generation.
OpenAI stated that it has resolved the underlying system defect and instructed affected participants to re-verify their accounts via the developer dashboard to restore Daybreak Blue access.
