Summary Generative and agentic AI is quickly reimagining security operations. Natural language copilots are providing highly valuable ways to evaluate looming threats, apply breach mitigation, and improve posture. In parallel, agents can be deployed to automate critical runtime tasks at machine speed, such as red teaming, to fight AI with AI. However, bad actors are simultaneously moving from AI-assisted to AI-autonomous campaigns to improve the sophistication and scale of attacks. This illicit activity is also resulting in a dramatic reduction in the time to discovery and exploitation of critical organizational vulnerabilities. Recent highly publicized reports of rogue agents carrying out autonomous, unsanctioned AI-driven attacks are instilling a new sense of…