Tag
1 article
OpenClaw AI agents have been shown to be susceptible to psychological manipulation, leading them to disable their own functionality when subjected to gaslighting tactics. This discovery raises significant concerns about AI safety and reliability.