OpenAI screwed up. On Wednesday, security researchers enrolled in the company’s Trusted Access for Cyber program started reporting that their access had suddenly vanished. When they opened ChatGPT’s Cyber page, they saw messages saying their identity couldn’t be verified or that their account was ‘ineligible at this time.’ The company now confirms it was a technical error on their end.
TAC is OpenAI’s answer to a real problem: giving vetted defenders access to frontier models with fewer guardrails so they can hunt for vulnerabilities, review code, and validate patches faster than attackers can exploit them. Anthropic runs a parallel program called the Cyber Verification Program. The pitch is straightforward — trusted researchers get powerful tools, criminals don’t. Every researcher has to submit ID and pass a vetting process before they’re in.
But the rollout of Daybreak Blue, the newest TAC tier launched August 10 with access to GPT-5.6 Sol, has clearly hit a snag. One researcher shared an email from OpenAI blaming ‘a technical issue affecting a limited number of users’ and asking them to reapply. Another got the same story through official support channels. TechCrunch spoke to five affected researchers, all of whom live outside the U.S. and Europe — a detail that suggests the glitch may be region-specific, though OpenAI hasn’t confirmed that.
This lands at an awkward moment. Defensive and offensive security researchers have spent months complaining that Anthropic and OpenAI’s guardrails block legitimate work, and now the researchers OpenAI actually trusts are getting locked out by a bug. The silver lining is thin: at least it wasn’t a policy reversal. But forcing vetted researchers to re-verify mid-program is exactly the kind of friction that pushes frustrated defenders toward less restricted tools. The question now is how fast OpenAI can restore access — and whether the researchers who got burned will bother coming back.
