AI tools are everywhere—but most are shockingly insecure by design.

In this fast-paced, demo-heavy session, we’ll hack real AI systems live—from prompt injections and data leaks to manipulating AI agents into executing unintended actions. You’ll see how attackers can exploit LLMs to bypass controls, extract sensitive information, and weaponize automation.

This talk is designed for both hackers and builders—showing not just how to break AI, but how to fix it.

Key Takeaways
Step-by-step AI hacking techniques
Live exploitation of LLM vulnerabilities
Common mistakes companies make with AI
Practical fixes to secure AI applications