Google has fixed two flaws in Vertex AI, its platform for custom development and deployment of large language models (LLMs), that could have allowed attackers to exfiltrate proprietary enterprise ...