# Security Experts Use AI to Develop Chrome Exploit in Record Time

*Published October 16, 2026*
*Source: [https://thehackernews.com/2026/10/researchers-used-public-openai-model-to.html](https://thehackernews.com/2026/10/researchers-used-public-openai-model-to.html)*

## Executive Summary

*This is a Premium edition. The Executive Summary is available to sec-news.ai members —*
*[read it here](https://www.sec-news.ai/news/security-experts-use-ai-to-develop-chrome-exploit-in-record-time) or [see plans](https://www.sec-news.ai/pricing).*

## Article

In an impressive demonstration of cybersecurity prowess, the team at XBOW, a security firm, successfully developed a full-chain exploit for Google Chrome in just six days. This remarkable feat was achieved by employing AI agents using the publicly available OpenAI GPT-5.6 Sol model. Google responded swiftly, confirmed the existence of three previously unknown bugs, and patched them, awarding XBOW a substantial $250,000 reward for their discovery.

The exploit crafted by XBOW proved to be a sophisticated chain of vulnerabilities, each aimed at bypassing Chrome's layered security measures. Chrome operates its web page code within two nested sandboxes designed to protect the underlying system from malicious activities. XBOW's chain managed to breach these defenses, enabling the execution of arbitrary code on a Windows machine without any user interaction beyond merely visiting a web page.

Google promptly addressed these vulnerabilities, ensuring user safety by fixing them in Chrome versions 152.0.7977.65 and 152.0.7977.82, released in late August and early September respectively. Although Chrome updates automatically, users are advised to restart their browsers to ensure they are protected by these patches.

The speed of this achievement is notable. XBOW initiated the challenge during a meeting in Malta in January and completed their first full-chain exploit over two months. However, their latest success during the Black Hat and DEF CON conferences in August only took six days, thanks to the integration of AI technologies. The project incurred an approximate cost of $40,000 in AI charges.

XBOW emphasizes the critical role of human expertise in guiding AI efforts. While the AI agents scoured Chrome's source code, tested hypotheses, and drafted initial exploit concepts, the researchers provided essential oversight. They selected targets, evaluated the AI's findings, and made strategic decisions to ensure the success of the exploit.

Before the vulnerabilities were patched, an estimated one billion Chrome users on Windows were at potential risk. This event underscores the importance of staying vigilant in the ever-evolving landscape of cybersecurity threats. The success of XBOW's work not only highlighted the potential of AI in cybersecurity research but also served as a wake-up call for defenders to adapt their strategies accordingly.