analiza aiqbench

OpenAI's automated research pipeline successfully identified and patched 948 Linux kernel privilege escalation vulnerabilities in a single run, demonstrating the practical application of advanced AI agents for large-scale security auditing. The study highlights both the efficiency of using autonomous models to scan millions of lines of code and the significant economic value these discoveries hold compared to traditional bounty programs. This work underscores a growing trend where AI is transitioning from content generation to critical infrastructure maintenance and threat remediation.