KEYNOTE: Attacking AI
Jason Haddix
Bug Bounty Village @ DEF CON 33 · Day 1 · Bug Bounty Village
In this keynote at Bug Bounty Village, Jason Haddix, a seasoned offensive security expert and leader at Arcane, delves into the rapidly evolving landscape of attacking **Generative AI**-powered applications. With the proliferation of AI features across web applications, APIs, and cloud services in the last three years, a significant new attack surface has emerged, demanding novel exploitation methodologies. Haddix's talk moves beyond traditional AI red teaming, which often focuses solely on manipulating AI models to generate harmful content, to encompass a holistic approach that targets the entire **ecosystem** of AI-assisted applications.
AI review
Haddix delivers a competent, practitioner-oriented survey of AI application security with real case study grounding — hardcoded keys in system prompts, RAG data leakage, over-scoped agent permissions — that's more useful than most vendor-led AI security content. The taxonomy (intent/technique/evasion) is a serviceable framework, and the case studies are concrete enough to be actionable. But for a Bug Bounty Village keynote, this is solid craft, not groundbreaking research — the individual findings are well-documented elsewhere, and the synthesis, while useful, won't surprise anyone who's…