AI researchers exposed a critical jailbreaking vulnerability allowing chatbots to generate harmful content with a single technique — a finding with serious implications for crypto projects betting on AI-powered infrastructure.
July 2, 2026
