Abstract
This article presents Dialogic Agent Theory for AI (DATA), a conceptual model that redefines AI-based communicative actors as discursive agents shaped through interaction, symbolic encoding, and institutional intent. Synthesized from peer-reviewed studies, DATA departs from mechanistic views of AI by stating that such agents acquire identity through tensions between corporate authorship, user projection, visual representation, and sociotechnical settings. Informed by Bakhtinian dialogism, organizational identity theory, and research in human–computer interaction, the theory proposes that AI agents are not static carriers of scripted roles but become recognizable through polyphonic emergence, temporal shifts, and interpretive co-authorship. The study employs an illustrative case analysis of Microsoft’s Tay chatbot to demonstrate the conceptual and practical implications of DATA. It shows how communicative failures are not limited to technical breakdowns but signal a misreading of dialogic interaction. DATA contributes to the analysis of artificial communicators in business, marketing, and public-facing contexts by showing how they represent an organization, shape user perception, and gradually take part in crafting its public voice.

This publication has 44 references indexed in Scilit: