AI Router Risks: Credential and Crypto Theft by LLM Intermediaries
University of California researchers have discovered that some third-party AI large language model (LLM) routers can pose security vulnerabilities that can lead to crypto theft. A paper measuring malicious intermediary…