Why Does AI Understand Some People Better Than Others? A Linguistic Perspective on Human-AI Communication and the Changing Human Mind

Abstract

This conceptual article examines a question that is important as generative artificial intelligence enters everyday communication: why does the same AI system appear to understand some people better than others? The article argues that AI performance in dialogue should not be viewed only as a property of the model itself. Human linguistic behavior, including vocabulary, discourse organization, explicitness, ambiguity, multilingualism, and communicative style, may also shape interaction outcomes. Drawing on recent research on linguistic convergence, personalization, human adaptation, and AI-assisted writing, the article proposes a two-way model of human-AI communication. In this model, people influence how AI interprets their messages, while repeated interaction with AI may gradually influence users’ language and cognitive strategies. Particular attention is given to multilingual speakers and the Uzbek-Russian-English context. The article does not claim that AI necessarily changes human thinking; instead, it proposes testable questions about linguistic adaptation, problem formulation, cognitive offloading, and transfer beyond AI-supported tasks. The proposed framework can guide future research.