Presenting the augmented self: extending sender effects of the hyperpersonal model in artificial intelligence-mediated communication

Artificial intelligence (AI) has been increasingly used to facilitate interpersonal communication. To provide a theoretical account of how the use of AI may affect interpersonal communication, we extend the hyperpersonal model by conceptualizing AI-mediated communication (AIMC) as computer-mediated communication (CMC) with an AI layer that participates in message production. Focusing on the sender side of the hyperpersonal model, namely selective self-presentation, we propose 2 AIMC affordances (i.e., co-creation and adaptability) and argue that these 2 affordances systematically reshape selective self-presentation by changing the production of both content (i.e., what to say) and language style (i.e., how to say it) of self-presentation. Using traditional CMC as the benchmark, we derive propositions specifying how self-presentation in AIMC differs from CMC, how these differences translate into impression and relational outcomes, and how they affect senders’ psychology. This article results in a theoretical framework called the Computational-Assisted Relational Exchange model, which advances a research program that investigates the changes introduced by AIMC relative to CMC and shows how classical CMC theories can be revisited, reexamined, and extended to account for the mechanisms involved.