Do We Conform to Machines? Social Influence in the Age of AI

People now take advice, drafts and judgments from AI systems many times a day, and classic social influence theory predicts they will sometimes yield to them. This article reviews how the work of Sherif, Asch, Deutsch and Gerard, and Cialdini on informational and normative influence applies to human–AI interaction. It synthesizes evidence on computers as social actors, algorithm appreciation and aversion, automation bias, robot-induced peer pressure, and the influence of large language models (LLMs) on opinions and moral judgments. Four mechanisms are proposed: perceived competence, authority cues, social presence, and cognitive effort reduction. Task ambiguity, self-efficacy, domain expertise, and culture are discussed as moderators. A qualitative study of ten adults, analysed with reflexive thematic analysis, is presented as a framework of eight superordinate themes. The article ends with design implications and a research agenda aimed at separating rational reliance from conformity.