In a major breakthrough for cybersecurity, Google CEO Sundar Pichai announced on Tuesday (July 15) that the company’s AI agent successfully identified and thwarted a cyber exploit before it could be deployed — a first-of-its-kind achievement for artificial intelligence in threat prevention.

“New from our security teams: Our AI agent helped us detect and foil an imminent exploit. We believe this is a first for an AI agent – definitely not the last – giving cybersecurity defenders new tools to stop threats before they’re widespread,” Pichai posted on X (formerly Twitter).

A new era in cybersecurity?

This marks a potential inflection point in cybersecurity, as AI shifts from passive defense — identifying threats post-breach — to proactive interdiction.

What’s next for AI agent

Google has not disclosed when Big Sleep was deployed or how long it has been operational. However, Pichai’s post suggests this is just the beginning of more AI-driven defense tools that will be used across Google’s ecosystem and offered to cloud clients.

As cyberattacks grow more frequent and damaging, the use of advanced AI may become standard across global IT defenses.

AI marks breakthrough in vulnerability detection

Big Sleep is an AI agent developed by Google DeepMind in collaboration with Google Project Zero, designed to autonomously detect unknown security vulnerabilities in software. In November 2024, it identified its first real-world vulnerability, marking a major milestone in AI-driven cybersecurity and highlighting its potential to detect threats before they can be exploited.

Since then, it has uncovered multiple real-world flaws, surpassing expectations and accelerating the field of AI-assisted vulnerability discovery. Most notably, the AI recently detected a critical flaw in SQLite (CVE-2025-6965), which had only been known to malicious actors. With support from Google Threat Intelligence, it was able to anticipate an imminent exploitation attempt and prevent it — marking what Google believes is the first successful use of AI to proactively stop a vulnerability from being exploited in the wild.

Beyond securing Google’s own ecosystem, it is now being used to bolster the security of widely used open-source software, significantly improving the broader internet’s safety. These AI-powered agents represent a major shift in cybersecurity, enabling human teams to focus on complex threats while dramatically expanding their reach and effectiveness.