The system builds on research showing AI can uncover serious software flaws across vast amounts of code, at a time when both hackers and defenders are turning to AI.
Anthropic's Claude AI autonomously discovered 500+ critical vulnerabilities in popular open-source software using only basic debugging tools.