Most security programs track AI risk as a list of separate findings: an exposed model endpoint, an over-permissioned IAM role, a storage bucket holding training data. Attackers don’t treat these as separate issues. They chain them.

At this executive dinner, you join security and cloud leaders to discuss how to tell theoretical AI threats apart from risks an attacker can exploit in your AWS environment, and how to decide which of those risks matter most to the business.

Horizon3 shares how NodeZero safely tests AWS AI workloads to show where weaknesses combine into real attack paths, and what each path puts at risk: data, models, or business operations.