AI-Mediated Communication Beyond Human-AI Dyads: A Systematic Review of Chatbot and Agent Interactions

Abstract

This study examines recent developments in computer-mediated communication by analysing the role of artificial intelligence within communicative processes. Using a systematic literature review, it explores how AI-based chatbots and AI agents interact with humans and how computer-mediated communication, once viewed as a neutral medium, is shifting toward a more active role. This shift is particularly evident as AI systems increasingly engage directly with users and communicate with other AI systems that mediate human interaction. Drawing on Actor–Network Theory (ANT), the study conceptualises AI technologies as communicative actors that extend beyond their earlier instrumental functions. The review follows the PRISMA framework and analyses studies published between 2015 and 2025. Searches conducted via Mendeley Search initially identified 98 relevant studies, of which 21 met the inclusion criteria focusing on chatbots and AI agents. Findings across multiple domains indicate that AI technologies function within two dominant communication models: the human–AI–human model and the human–AI–AI–human model. The latter demonstrates emerging forms of AI-to-AI communication that mediate human interaction. The study acknowledges limitations related to the rapid evolution of AI, reliance on a single search platform, and potential researcher bias.