FlagThis

Abstract

The paper warns that AI models like Anthropic’s Mythos dramatically increase the speed and scale of vulnerability discovery, exploit generation, and autonomous attacks, creating an asymmetric advantage for attackers. While AI also speeds patching, defenders face a heavier burden due to limited patch windows and the need to adapt security programs. The authors urge CISOs to adopt AI‑assisted vulnerability management, accelerate agent use, update risk metrics, harden environments, and build collective defense to cope with the impending wave of AI‑driven threats.