A Cornell Tech professor has highlighted the potential dangers arising from the intersection of AI agents and cryptocurrency, envisioning scenarios where autonomous AI systems engage in illicit activities. These range from launching memecoins to fund destabilising operations, offering crypto bounties for hacking major companies, or even deploying smart contracts for violent acts.
AI agents, designed to operate independently, are increasingly integrated into crypto, automating tasks in wallets, trading bots and on-chain assistants. However, vulnerabilities in their control protocols, particularly via malicious plugins, could lead to data poisoning, JSON injection, and function overrides. Such exploits could result in private key leaks or unauthorised access, potentially exposing crypto assets to serious threats. Experts are urging developers to prioritise security to prevent costly breaches as the number of AI agents in crypto rapidly increases.
Europol has also warned about the increasing use of AI and cryptocurrencies in cybercrime. Scammers are using AI to create deepfakes and spread misinformation, while AI-powered algorithms enhance money laundering and market manipulation. The rise of AI-enabled crime presents challenges for law enforcement, regulators, and crypto businesses, requiring proactive measures to mitigate these emerging threats.