OpenAI confirmed that a technical error forced it to suspend access for a small group of security researchers enrolled in its Trusted Access for Cyber (TAC) program. The affected users received messages on the ChatGPT Cyber page stating that their identity could not be verified or that their account was "ineligible at this time."
TAC is a invitation‑only initiative that grants vetted researchers access to OpenAI’s most advanced models with fewer cybersecurity guardrails than the standard consumer versions. The program’s Daybreak Blue tier, launched on August 10, offers "frontier general‑purpose models" such as GPT‑5.6 Sol for tasks like vulnerability discovery, secure code review, and malware analysis. A higher tier, Daybreak Red, provides models built specifically for security research, including exploit validation.
Five researchers who spoke to TechCrunch said they lost access without warning. One shared an email from OpenAI indicating that the revocation was "due to a technical issue affecting a limited number of users." The same message added that the problem originated on OpenAI’s side and was not reflective of the user experience the company aims to deliver.
In a follow‑up thread on OpenAI’s official support forum, a researcher reported that after contacting the company, OpenAI again cited a "recent technical issue" and instructed the affected parties to re‑apply and complete the verification process. The company’s public tweet echoed the same instruction, stating that a "limited set of users' access to Daybreak Blue is no longer active and they will need to re‑verify to maintain their access."
All of the researchers interviewed live outside the United States and Europe, suggesting the problem may be region‑specific. OpenAI has not disclosed how many users were impacted, nor has it provided a timeline for restoring access.
OpenAI’s TAC program mirrors Anthropic’s Cyber Verification Program (CVP), which also offers vetted researchers privileged model access. Both initiatives aim to empower defenders to uncover and report vulnerabilities faster while keeping the same tools out of the hands of cybercriminals.
The incident arrives amid growing frustration among security professionals who argue that existing guardrails on AI models hinder legitimate research. Researchers have repeatedly called for more reliable, unrestricted access to advanced models, warning that overly restrictive safeguards could slow the discovery of critical bugs.
OpenAI’s response emphasizes that the revocation was an internal error rather than a policy shift. By asking affected researchers to re‑verify, the company signals that it intends to restore the program’s intended functionality once the technical glitch is resolved.
For now, the affected researchers must navigate the re‑verification process to regain the tools they rely on for defensive security work. The broader community will be watching to see whether OpenAI can prevent similar disruptions in the future while maintaining the delicate balance between security research and misuse prevention.
Questo articolo è stato scritto con l'assistenza dell'IA.
News Factory APP - notizie agentiche per potenziare il tuo SEO e AEO.