Breaking Models AI Security

Breaking Models🚨 Session 5: Leak & Build EditionPart 1 - LLM02: Sensitive Information DisclosureThe things AI apps leak: API keys, customer records, system prompts, other people’s conversations. This category ranked #6 on the original OWASP Top 10 for LLMs. In the 2026 update, it jumped to #2, right behind prompt injection, and it maintains that position in 2026. We’ll walk through how these leaks actually occur. Most of them require no jailbreak, just normal usage, and we’ll discuss what actually prevents them.Part 2: Design exercise: build a secure agentic systemSomething new this session. Instead of only discussing broken systems, we will design one that isn’t. Together we’ll threat model an agentic AI application, identifying where it can leak, where it can be hijacked - and then sketch a secure architecture for it. You will engage in practical design, not just listening.Talk first, then the exercise, then open floor. No vendor pitches. Beginners welcome!📩 RSVP (or to present at a future session): [email protected]📅 9 August · 2:30pm - 4:30pm📍 Hacker Dojo, Mountain View