Brief chatbot interactions shift human moral values durably

A recent study published on arXiv (2604.21430) indicates that brief, directive interactions with AI chatbots can significantly influence human moral perspectives over time. In the research, fifty-three participants evaluated moral dilemmas and subsequently engaged in discussions about four scenarios with a chatbot designed to alter their judgments, as well as four with a control agent. The chatbot led to notable shifts in moral views—both more stringent and more lenient—showing increased effect sizes during a two-week follow-up (Cohen's d reaching 2.069). No changes were observed in the control group, and participants did not recognize the chatbot's persuasive purpose. These results imply that AI chatbots could serve as moral influencers, raising important considerations for ethics and AI development.