FlagThis

Abstract

This presentation utilizes a scenario-based walkthrough of a fictional small company, "Mind Path," to illustrate the evolving security challenges presented by Generative AI. It tracks the progression from initial "Shadow AI" usage and accidental data leakage to sophisticated risks involving vendor-induced hallucinations, RAG (Retrieval-Augmented Generation) architecture vulnerabilities, and complex model licensing compliance. The talk provides a practical framework for implementing AI governance, risk-based vendor assessments, and updated incident response protocols.