Researchers say OpenAI revoked their access to limited cyber program

OpenAI suddenly revoked researchers' access to a limited cybersecurity program called Trusted Access for Cyber (TAC), which gives vetted security researchers access to advanced AI models with reduced safety restrictions for legitimate defensive work. OpenAI confirmed the revocation was caused by a technical error and asked affected researchers to reapply and reverify their credentials. The program exists to allow trusted defenders to find bugs and vulnerabilities faster while preventing malicious actors from accessing these powerful models for harmful purposes. All the researchers affected reported living outside the U.S. and Europe, suggesting the issue may have been limited to certain regions.