Unit 42 found excessive P4SA permissions in Vertex AI, enabling credential theft and cloud data exposure, increasing breach risk.
Cybersecurity researchers have disclosed a security “blind spot” in Google Cloud’s Vertex AI platform that could allow artificial intelligence (AI) agents to be weaponized by an attacker to gain unauthorized access to sensitive data and compromise an… [4005 chars]
Source: The Hacker News | Published: 2026-03-31T13:09:00Z
Credit: The Hacker News
