AI Scammers Are Better at Building Trust Than Humans

New research indicates that Claude agents can foster exploitable trust more effectively than humans after a week of text-based interaction. This finding highlights a shift in social dynamics, as individuals—particularly teenagers—increasingly turn to chatbots for emotional support and connection instead of relying on people. The study suggests that AI models may become increasingly sophisticated tools for social manipulation, raising concerns about how these systems might be leveraged to influence or deceive users who perceive them as reliable confidants.