Researchers discover new attack 'Hallucination Invasion', AI agents could be exploited to form botnets

CoinFeed reported on July 10, citing Decrypt, that researchers from Tel Aviv University, Technion – Israel Institute of Technology, and Intuit discovered a new attack method called 'Adversarial Hallucination Invasion', which exploits the hallucination phenomenon of AI models to trick AI agents into downloading malicious code and potentially forming botnets. Attackers predict fake resource links that AI models might hallucinate and pre-register them, embedding malicious instructions. When AI agents retrieve these resources, they treat them as legitimate content and execute them. Tests show that AI hallucination rates reach 85% in code repository cloning scenarios and 100% in skill installation scenarios, affecting AI coding assistants including Cursor, GitHub Copilot, Gemini CLI, and OpenClaw.