Can You Trust Your LLM? A Practical Guide to AI Security

LLMs can write code, search documents, call APIs, and automate entire workflows—but what happens when they get access to sensitive data or make the wrong decision?

In this talk, we'll look at the practical security risks that come with using and building LLM-powered applications. We'll cover how to use hosted AI tools without unnecessarily exposing private information, when local and self-hosted models can provide stronger privacy guarantees, and how to design secure applications around models that shouldn't be blindly trusted.

We'll explore topics including prompt injection, data leakage, RAG security, secrets management, tool permissions, authorization, least privilege, and limiting the blast radius of AI agents.

The goal isn't to make LLMs perfectly trustworthy. It's to build systems that stay secure even when the model isn't.

Whether you're experimenting with ChatGPT, building RAG applications, or developing autonomous agents, you'll leave with a practical framework for thinking about AI security from prompt to production.

LOGISTICS AND PARKING:

The talk starts at 7:00 PM. The first half hour is reserved for everyone to get set up and mingle. Free pizza and drinks!

The cheapest parking option is to find street parking, which will only cost you a few bucks. Otherwise, park in the nearby veteran's museum lot for $8. It's highly recommended you avoid the nearby $15 garage parking.