This literature review examines how contemporary scholarship theorizes artificial intelligence as an institutional actor, focusing on whether and how AI systems gain legitimacy, authority, and norm-shaping capacity within organizational and societal fields. Drawing on an AI-assisted discovery and screening workflow, the review identifies 1,093 candidate publications and narrows them to 50 high-relevance papers spanning media and journalism studies, organizational and management research, sociology, law, public governance, and higher education. The synthesis shows broad recognition that algorithms can function institutionally by structuring behavior and decision-making, yet the field remains conceptually fragmented and empirically underdeveloped. Most empirical studies analyze institutional adaptation and governance responses to AI rather than tracing processes through which AI systems themselves become embedded as durable rule-makers. The most promising advances come from sociotechnical and distributed-agency approaches, including actor-network theory and hybrid human–algorithm configurations, but these frameworks often leave ontological questions of “actorhood” unresolved. The review concludes that a mature, evidence-based theory of AI as institutional actors is still lacking, and it outlines priority research directions for empirically studying legitimation, institutionalization indicators, and cross-disciplinary integration.
Gerin Trautenberger (Sat,) studied this question.