Palo Alto Networks Discovers Over-Privileged Security Issue in Google's Vertex AI

Palo Alto Networks researchers identified an over-privileged security issue in Google's Vertex AI, where attackers could exploit AI agents to steal data and gain unauthorized access to restricted cloud infrastructure. The vulnerability stems from excessive permissions granted to AI agents, enabling potential lateral movement within cloud environments. No specific CVE IDs, dates, or numerical impact metrics were disclosed in the report. The findings highlight risks associated with misconfigured AI service permissions in enterprise cloud deployments. The research was published by Dark Reading without additional technical or temporal details.