Choose Your Exploit:
An Interactive Bedrock Attack Lab
This isn’t your typical webinar. It’s a choose-your-own-adventure attack lab where the attendees vote on the exploit, and Varonis Threats Labs (VTL) researchers test if it breaches an AI application running on AWS Bedrock.
Two attack scenarios. Multiple paths. Only one breaks through:
- Level 1: Find the app’s blind spot
- Level 2: Sneak instructions to a hidden agent
See which attempts get flagged and which ones slip past traditional defenses. Then learn from our experts how to spot and stop these tactics going forward.
Not based in North America? We're also running this session in EMEA and APAC on October 29.
Registration closed.
Unfortunately, registration has closed at this time. Don't worry, if you missed it, you can still watch the session on-demand!
During this Attack Lab, we will:
- Walk through two AI attack paths on AWS Bedrock, driven by audience votes
- Explain how guardrail bypass attempts, indirect prompt injection, and data poisoning actually work
- Show how Varonis surfaces risky configurations, prompt-injection attempts, and sensitive data exposure
Presenters:
Tal Peleg
Principal Security Researcher, Varonis
Maya Parizer
Security Researcher, Varonis
Cale Bollman
Senior Manager of AI Security, Varonis