Google’s AI-powered vulnerability researcher, Big Sleep, has successfully identified 20 security flaws in widely used open-source software, including tools like FFmpeg and ImageMagick. This initiative, developed by Google’s DeepMind and Project Zero teams, emphasizes augmenting human security researchers rather than replacing them, allowing for thousands of tests to be conducted rapidly. The AI operates by simulating malicious actions to probe code and network services for potential exploits, with human experts reviewing findings before submission. This advancement showcases the growing role of AI in enhancing cybersecurity efforts, reflecting a proactive approach to safeguarding software used globally.