HUMAN–AI INTERACTION IN HIGHER EDUCATION: HOW STUDENTS ADDRESS, PERSONIFY, AND EVALUATE AI LANGUAGE MODELS

Abstract

The aim of this article is to examine how Ukrainian university students employ AI language models both as practical tools and as communicative partners, reflecting broader cultural patterns of interaction. The material consists of survey data from 445 respondents representing different fields of study. The methodological design combines quantitative analysis of the frequency and contexts of AI use with qualitative examination of individual communicative practices, thereby enabling the identification of both general tendencies and individual strategies.

The findings indicate that ChatGPT is the most frequently employed tool, used for summarising texts, generating ideas, retrieving factual information, producing alternative formulations, and assisting in the fulfilment of academic assignments. Alongside these functional tasks, many respondents employ cultural and social conventions characteristic of human communication in their interactions with AI: they assign names to the system, use gendered pronouns, or use politeness strategies such as greetings and expressions of thanks. These practices show that the system is perceived not only as a technical instrument but also as a quasi-partner in discourse.

The survey shows that students approach the material produced by AI with caution: they recognise its convenience for routine operations, yet at the same time note recurring shortcomings, including factual inaccuracies, repetitive phrasing, and explanations perceived as overly simplistic. The analysis indicates that age affects the way students formulate their queries. Younger respondents often formulate brief and utilitarian requests reflecting a pragmatic orientation toward the task, with little attention to style or expression. Older participants, by contrast, more frequently produce expanded queries in which the content of the request is combined with conventional markers of politeness such as greetings or expressions of thanks.

This tendency demonstrates that communicative style in interaction with AI varies according to both functional purpose and sociocultural contexts. The analysis thus highlights AI's dual function in student communication: it serves as a technical instrument and a partner in interaction. The contribution of this research lies in outlining how these patterns can inform the development of pedagogical strategies that support reflective and responsible use of AI in higher education.