MODELEAK: PRIVILEGE ESCALATION TO LLM MODEL EXFILTRATION IN VERTEX AI

Summary from the TLDR InfoSec newsletter

Palo Alto Networks discovered two vulnerabilities in Google's Vertex AI platform: privilege escalation through custom jobs and model exfiltration via malicious models that could expose sensitive AI models and data. This post dives into the vulnerabilities and how they were identified.