OWASP Copenhagen Chapter

Want to learn about hacking LLMs?
Then you're in luck!

We're so lucky to have Kevin Joensen from Baldur Security coming to tell us about the recent challenge he made where he invited our community to hack the LLM he created for the chance to win tickets for BSides Copenhagen.
He'll talk about the basics of hacking LLMs, how he made the BSides CTF, how to defeat it - and how you can (try) to hack the new LLM CTF he created just for us!

Here's the schedule and description he sent us:

17:00: Welcome
17:10: Talk about common vulnerabilities and misconfigurations seen in LLM/AI enabled systems.
18:00: Pizza!
18:30: LLM CTF

We’ll start with an in-depth talk on vulnerabilities in Large Language Models (LLMs) and AI-powered systems, exploring topics like prompt injection, data leakage, model manipulation, and real-world attack scenarios inspired by cases from Microsoft and others. Following the talk, you’ll get to apply what you’ve learned in a hands-on CTF with AI-focused challenges. These range from bypassing chatbot safeguards to exploiting misconfigured AI systems. This event is perfect for anyone looking to better understand, identify, and mitigate emerging AI security risks.