A recent project by XBOW, a security firm, has demonstrated the potent capabilities of AI in exploit development, successfully chaining three previously unknown vulnerabilities in Google Chrome into a full-chain exploit. This attack, which can execute code on a Windows system by merely visiting a compromised web page, was developed in just six days using OpenAI's publicly available GPT-5.6 Sol model. Following the discovery, Google promptly confirmed the vulnerabilities, patched them, and rewarded XBOW with a substantial bounty.
The implications of this development are significant for businesses, particularly in understanding the evolving threat landscape where AI can accelerate malicious activities. Organizations must prioritize robust cybersecurity measures, including regular updates and vulnerability assessments, to protect against such rapid exploit creation. This incident underscores the intersection of AI and cybersecurity, reminding stakeholders that while AI can enhance defense mechanisms, it can also empower attackers, necessitating a proactive approach to risk management and threat mitigation.
---
*Originally reported by [The Hacker News](https://thehackernews.com/2026/10/researchers-used-public-openai-model-to.html)*