The HIPUD model

Abstract #

Gen-AI chatbots such as ChatGPT or Claude are becoming central tools for everyday information use. We argue that this shift is epistemic, not merely technological. Rather than simply mediating access to existing content, chatbots synthesize it into novel informational artifacts, changing how people encounter, process, and use information. To capture this shift, we propose the HIPUD model, which characterizes Gen-AI chatbots as information environments that are Hyper-personalized, Interactive, Probabilistic, Universal, and De-contextualized. The model organizes these characteristics in three layers: features of the system (probabilistic, universal), default features of the output (hyper-personalized, de-contextualized), and interactivity as the affordance through which users can partly adjust these defaults. We discuss broader transformations resulting from Gen-AI chatbot use and illustrate the model’s analytical leverage with examples from political, health, and consumer communication research. Overall, the model equips future research with a shared vocabulary and analytical lens for understanding information use in Gen-AI environments.

Note: This is a second, revised version of the preprint. The first version is available here: https://doi.org/10.31235/osf.io/45qnm_v1

Download

Cite