AI agents are accelerating vulnerability discovery. Here’s how AppSec teams must adapt.

The New Stack

by Josh Lemos

February 19, 2026

AI-Generated Deep Dive Summary

AI-powered tools are revolutionizing vulnerability discovery by making it faster and more scalable than ever before. Traditional methods, inspired by Linus’s Law, relied on human eyeballs to spot bugs, but AI has supercharged this process. Autonomous systems like XBOW have demonstrated remarkable efficiency, identifying over 1,060 vulnerabilities in just 90 days—a feat unmatched by most human teams. This shift is critical for AppSec professionals, as it raises the stakes: will organizations find and fix vulnerabilities before malicious actors do? AI red teaming tools are not only surpassing human capabilities but also operating at machine scale, testing entire attack surfaces simultaneously and methodically. The rise of AI-driven threat modeling further underscores its transformative potential. Enterprises like JPMorgan Chase are integrating generative AI into their security processes through systems like Auspex, which condenses weeks or even months of traditional threat modeling work into mere minutes. By embedding expert knowledge and best practices into AI prompts, these tools generate highly accurate threat matrices, enabling developers to address risks more effectively. This approach streamlines the development lifecycle, reducing inefficiencies that often lead to security debt and unpatched vulnerabilities. For DevOps teams, the implications are profound. AI not only accelerates vulnerability discovery but also shifts focus toward proactive risk management. By leveraging AI-driven intelligence, AppSec teams can systematically identify patterns across

Verticals

devopscloud

Originally published on The New Stack on 2/19/2026