Google’s AI Tools for Actively Hunting and Fixing Critical Cyber Vulnerabilities

Google’s AI Innovations in Cybersecurity

Since its inception last year, Google’s AI agent “Big Sleep”has made significant strides in identifying and addressing security vulnerabilities. Most notably, it recently uncovered CVE-2025-6965, a serious SQLite vulnerability that was actively being exploited by cybercriminals.

Google claims this is a groundbreaking moment as it marks the first instance where an AI agent has directly intervened to mitigate an ongoing exploit.According to Google, Big Sleep has also been utilized to bolster the security measures of critical open-source projects. This proactive approach to vulnerability management enhances the toolkit available for cybersecurity professionals combating evolving threats.

Enhancements in Digital Forensics

In addition to Big Sleep’s achievements, Google has unveiled new AI-driven capabilities intended to strengthen cybersecurity defenses. One such innovation is Timesketch, an open-source digital forensics platform being upgraded with functionalities supported by Sec-Gemini. This enhancement aims to automate the initial stages of forensic investigations, thereby accelerating incident response times.

Upcoming Innovations and Collaborations

Furthering its commitment to cybersecurity, Google plans to showcase FACADE (Fast and Accurate Contextual Anomaly Detection) at the upcoming Black Hat conference in August. Internally utilized since 2018, this tool has processed billions of daily security events, enabling the identification of internal threats without reliance on historical attack data.

Additionally, Google announced a contribution of data from its Secure AI Framework (SAIF) to support the Coalition for Secure AI (CoSAI), an initiative it helped establish. This data donation will facilitate CoSAI’s ongoing efforts in developing agentic AI, enhancing cyber defense capabilities, and securing software supply chains.

AI Cyber Challenge at DEF CON 33

Looking ahead, Google will also announce the winners of the AI Cyber Challenge in collaboration with DARPA during DEF CON 33 next month. Participants will present innovative AI solutions designed to uncover and remediate vulnerabilities within major open-source projects.

Through these pioneering efforts, Google hopes to contribute significantly to creating a safer digital environment for users globally.

Image via Depositphotos.com

Source&Images