Ignore All Previous Instructions: Hacking AI chatbots and Agents with Cameron Cottam

AI chatbots and agents are being added to everything, and many of them can be talked into misbehaving.

This talk explains how LLMs and AI agents are built, the main ways attackers abuse them, and why prompt injection is so hard to fix. Then it’s your turn: a live CTF where you trick a chatbot and an AI agent into leaking flags. Bring a laptop. No AI experience needed.