Google is enhancing its AI-driven cybersecurity tools and partnerships, showcasing innovations at Black Hat USA and DEF CON 33. Notable is Big Sleep, an AI agent that identifies unknown vulnerabilities before exploitation. By November 2024, it uncovered a major threat, marking a breakthrough for AI in security. Recent findings, including CVE-2025-6965 in SQLite, highlight proactive measures in protecting critical infrastructures. Google is also upgrading Timesketch for forensic analysis and demonstrating FACADE's advanced anomaly detection capabilities, which operate without relying on historical data.
Big Sleep is an AI agent developed by DeepMind and Google Project Zero that identifies unknown software vulnerabilities before they are exploited, achieving significant milestones in AI-assisted security.
The integration of Sec-Gemini into Timesketch will enable AI to conduct preliminary forensic investigations, reducing the manual workload for analysts and allowing them to focus on prioritizing threats.
FA CADE uses contrastive learning to detect insider threats without depending on historical attack data, marking a departure from traditional detection systems and enhancing threat identification.
Big Sleep's discovery of CVE-2025-6965 in SQLite is the first known instance of an AI agent proactively stopping an exploit before deployment, showcasing the capabilities of predictive AI in cybersecurity.
Collection
[
|
...
]