Google's AI agent Big Sleep detects first major cybersecurity vulnerability

Besides early detection, Google's Big Sleep AI agent was able to find a fix and prevent the threat actors from exploiting it beforehand.

Google's Android, Play Store, ChromeOS and other products are used by billions of people around the world.

It is imperative for the American company to offer a safe environment for users and also prevent them from falling prey to cybercriminals online.

Whenever Google detects any vulnerabilities in its products, it rolls out security patch to fix them.

For early detection of security issues, Google DeepMind and the company's top security watchdog, Google Project Zero, in 2014 announced the launch of Big Sleep, the company's first-ever generative Artificial Intelligence (gen AI)-powered agent.

By the end of 2024, it was able to make the first discovery of a security vulnerability. Though it was not a major breakthrough, but showed the potential of Google's AI agent capabilities.

Now, the search engine giant has revealed that the Big Sleep was able to decipher an intel tip from Google Threat Intelligence to detect SQLite vulnerability (CVE-2025-6965), understood to be a critical security flaw in one of Google's products. It was known only to threat actors and was at risk of being exploited.

Before they could misuse the vulnerability, Big Sleep was able to find a fix and prevent the threat actors from exploiting it beforehand.

Through the combination of threat intelligence and Big Sleep, Google was able to predict that a vulnerability was imminently going to be used, and we were able to cut it off beforehand. We believe this is the first time an AI agent has been used to directly foil efforts to exploit a vulnerability.

Kent Walker, President of Global Affairs, Google & Alphabet

Google has announced that it will be offering a demo of its advanced AI-powered cybersecurity agent and products such as Timesketch and FACADE (Fast and Accurate Contextual Anomaly Detection early next month.

For the uninitiated, Timesketch is Google’s open-source collaborative digital forensics platform, with agentic capabilities. It can assist cyber forensic experts with incident response by using AI to automatically perform the initial forensic investigation.

FACADE is being used to screen daily security events across Google to identify internal threats.

In August, Google and Airbus, in collaboration with the US military's Defense Advanced Research Projects Agency (DARPA), will host the final round of the two-year AI Cyber Challenge (AIxCC) at DEF CON 33 event (August 7-10, 2025) in Las Vegas, California.

Get the latest news on new launches, gadget reviews, apps, cybersecurity, and more on personal technology only on DH Tech.