McKinsey’s internal AI Platform called Lilli got hacked by another AI in 2 hours because they failed to adopt the most basic of security practices: exposed endpoints, unauthenticated APIs, and SQL injection, ultimately leading to leaked chat messages, documents, user accounts, AI systems, system prompts, and RAG chunks. This is another case of where AI really enables things at scale, including finding and exploiting security vulnerabilities. So companies and people have to be real diligent and get on top of securing their assets —— Follow @cyberjasbytes to stay safe in the age of AI —— All media from Codewall #mckinsey #big3 #mbb mckinsey hack mckinsey ai what happened with mckinsey database exfiltration leaked ai ai hack ai research ai security tech news vibe coding security how to use ai securely building with ai builder consulting firm