Google’s Big Sleep AI Foils Live Zero-Day Exploit in SQLite (CVE-2025-6965)

In a demonstration of artificial intelligence applied to cybersecurity, Google has revealed that its AI agent, Big Sleep, has successfully identified and neutralized a critical vulnerability before it could be exploited in the wild. This marks a new era of AI-powered proactive defense, where machines not only detect threats but preemptively stop them.

“By November 2024, Big Sleep was able to find its first real-world security vulnerability, showing the immense potential of AI to plug security holes before they impact users,” Google stated in its blog post.

Originally announced last year by Google DeepMind and Google Project Zero, Big Sleep was designed to autonomously hunt for zero-day vulnerabilities across complex codebases. But its most recent win may be the most significant yet.

According to Google, Big Sleep identified CVE-2025-6965, a memory corruption flaw in SQLite, one of the world’s most widely used embedded databases. The bug, which existed in SQLite versions prior to 3.50.2, could allow attackers to corrupt memory by exceeding the number of aggregate terms beyond available columns—a potentially serious vulnerability.

“The Big Sleep agent discovered an SQLite vulnerability — a critical security flaw, and one that was known only to threat actors and was at risk of being exploited.”

Google’s Threat Intelligence team had already suspected this vulnerability was being targeted, and Big Sleep was able to confirm and locate it before it was exploited at scale.

“We believe this is the first time an AI agent has been used to directly foil efforts to exploit a vulnerability in the wild.”

This marks a major milestone in cybersecurity history: AI not only detected but disrupted an impending cyberattack.

While Big Sleep helps protect Google’s products, its influence now extends beyond proprietary systems. Google is deploying Big Sleep to scan and secure open-source projects, helping safeguard the wider internet ecosystem.

“These AI advances don’t just help secure Google’s products. Big Sleep is also being deployed to help improve the security of widely used open-source projects — a major win for ensuring faster, more effective security across the internet more broadly.”

Despite the power of Big Sleep, Google emphasized the importance of safe, responsible AI deployment. In its latest white paper, the company outlines guardrails like human oversight, privacy safeguards, and transparency.

“When deployed according to secure-by-design principles, agents can give defenders an edge like no other tool that came before them.”

Big Sleep’s operations are governed by strict industry-standard disclosure processes, and all its findings are publicly trackable via Google’s issue tracker.

Related Posts:

• Warning: “Sleeper Agent” Chrome Extensions Infect 1.5 Million Users!
• GPU-Powered Evasion: Unpacking the Sophisticated CoffeeLoader Malware
• Sophisticated NFT Scam Campaign Exposed: Over 100 Projects Targeted
• Mirai Botnet Unleashes Record-Breaking DDoS Attack, Cloudflare Thwarts Threat
• Europol & Microsoft Lead Global Takedown of Lumma Stealer, World’s Largest Infostealer