AI finds unexpected bugs in cybersecurity challenge

During DARPA's Artificial Intelligence Cyber Challenge (AIxCC), advanced AI systems designed to find vulnerabilities in software code exceeded expectations. These AI tools scanned millions of lines of code, successfully identifying numerous artificial flaws introduced by DARPA. Remarkably, the AI also uncovered over a dozen previously unknown bugs that were not intentionally planted. This development highlights the growing capability of AI in uncovering complex security issues, even those not anticipated by human experts, suggesting a significant leap in automated cybersecurity analysis.