Google announced it took an incredible leap in cyber defense using artificial intelligence. One of the world’s trendiest technologies was reportedly capable of stopping one of the globe’s biggest threats. According to the tech giant, a new AI agent successfully stopped a cyberattack before any information was breached.
Google said that AI provides “unprecedented opportunity” to help security defenders.
AI Agent Identifies and Deflects Cyber Threat
The breakthrough in cybersecurity came from Google DeepMind and Google Project Zero’s “Big Sleep” AI agent.
Google announced Big Sleep last year as an agent designed to search for unknown security threats in software. According to the press release, Big Sleep successfully identified its first security vulnerability in November 2024. Its ability to target the threat showed AI’s powerful potential in addressing a security flaw before it impacts a user.
The Google team says Big Sleep has exceeded their expectations since then and has discovered multiple real-world vulnerabilities.
Explore Tomorrow's World From Your Inbox
Get the latest science, technology, and sustainability content delivered to your inbox.
I understand that by providing my email address, I agree to receive emails from Tomorrow's World Today. I understand that I may opt out of receiving such communications at any time.
Based on information from Google Threat Intelligence, its most notable detection was an SQLite vulnerability. It found a “critical security flaw” known only to threat actors and at risk of exploitation.
With combined efforts from threat intelligence and Big Sleep, Google successfully predicted the imminent exploration and handled the threat. The company said, “We believe this is the first time an AI agent has been used to directly foil efforts to exploit a vulnerability in the wild.”
A Game Changer For Cybersecurity
According to Google, the Big Sleep AI agent is also deployed outside the company’s products. For example, it is being used for widely used open-source projects. Google views cybersecurity agents as “game changers.” They believe they help teams concentrate on high-complexity threats and expand their impact.
Google emphasized that the work was done “safely and responsibly.”
The company outlines its approach to building AI agents that protect privacy and avoid rogue actions. It believes AI defenders offer an “unparalleled advantage” if deployed correctly and following “secure-by-design principles.” Google promises to share developing AI insights and findings through the standard disclosure process.
In addition, Google plans to demo new AI capabilities for defenders that sift through vast amounts of data to identify threats.
