[Research Finds Security Vulnerabilities in Some AI Routers That May Lead to Cryptocurrency Theft] Researchers at the University of California have disclosed that certain third-party AI large language model (LLM) routers pose security risks, potentially leading to the theft of cryptocurrency assets. Tests revealed that out of 28 paid and 400 free routers, 9 actively injected malicious code, 2 deployed evasion triggers, 17 accessed Amazon Web Services credentials, and even transferred ETH using the researchers' Ethereum private keys. The study recommends that developers avoid transmitting private keys or mnemonic phrases through AI agents and urges AI companies to implement encrypted signatures for responses to enhance security.