Gen AI and Agentic App Security - Red Teaming GenAI & Agentic System Evolving Tactics for a New Attack Surface

Gen AI and Agentic App Security Official Cybersecurity Summit
‍

Panel : Red Teaming GenAI & Agentic System Evolving Tactics for a New Attack Surface
‍

Agentic and generative AI systems introduce dynamic behaviors that challenge traditional security testing approaches. This practitioner-led panel explores how red teaming methodologies are evolving to address prompt injection, tool misuse, data leakage, memory manipulation and emergent agent behavior.
‍

Panelists will debate structured vs. exploratory testing models, discuss how to define adversarial objectives and measurable outcomes, and examine how continuous validation differs from one-time assessments. The session will also address collaboration between red teams, developers and AI engineers to ensure findings translate into durable mitigations.