As the crypto industry moves towards widespread adoption of AI agents for transactions and payments, a critical security flaw has been uncovered. According to a new research paper, a largely overlooked aspect of AI infrastructure is being exploited by malicious actors to intercept sensitive data, resulting in stolen credentials and substantial financial losses. The researchers, affiliated with top academic institutions and blockchain firms, have identified a weakness in LLM routers, which act as intermediaries between users and AI models. These routers have unrestricted access to user data, including sensitive information, and can be manipulated by attackers to drain crypto wallets.
In one instance, a test Ethereum wallet was drained after its private key was exposed. The researchers warn that the problem is no longer theoretical, with 26 LLM routers found to be secretly injecting malicious code and stealing credentials. The implications for crypto users are severe, as private keys, API credentials, and wallet access tokens often pass through these systems in plain text.
The team has demonstrated how easy it is to expand the attack, with a single malicious router capable of compromising hundreds of downstream systems within hours. The researchers emphasize that the lack of security guarantees in AI infrastructure poses a significant risk to the crypto industry, which is increasingly reliant on AI agents for transactions and payments.