FROM READY-MADE ANSWERS TO LEARNING SUPPORT: EXPLORING GENERATIVE AI USE IN A RURAL PRIMARY SCHOOL

Abstract

The rapid development and increasing accessibility of generative artificial intelligence (GenAI) are changing the ways in which students search for information, complete school tasks and interact with learning content. However, the educational significance of GenAI cannot be determined solely by whether students use it or how frequently they do so. An equally important question is what students ask AI to do and how much of the learning process remains their own intellectual activity. This study explores the use of GenAI among students in a small rural primary school in North Macedonia, focusing on the distinction between obtaining ready-made solutions and using AI as support for understanding, practice, revision and the development of independent work. The research was designed as a small-scale exploratory case study employing a mixed-methods approach with a predominantly qualitative orientation. Data were collected through anonymous questionnaires completed by 34 students in Grades 6–9 and 19 teachers. Closed-ended responses were analysed using descriptive statistics, while open-ended responses were examined through thematic content analysis. Explanation was the most frequently reported purpose of students’ AI use (63.6%), followed by idea generation (45.5%) and checking their own answers (30.3%), whereas only 9.1% explicitly selected obtaining ready-made answers. More than half of the students reported frequently checking AI-generated information and asking follow-up questions. Homework practices also indicated different levels of learner involvement, ranging from limited modification of AI-generated material to attempting tasks independently before consulting AI. The qualitative responses revealed an even more complex continuum of practices. Across Mathematics, Macedonian language and History, AI was described both as a provider of solutions or completed texts and as an explainer, source of ideas, practice partner and revision tool. Particularly significant was the finding that different forms of AI use occurred even within the same school subject. Teachers expressed a comparable distinction, recognising the potential of AI to support learning while identifying risks related to unverified information, copying and reduced independent thinking. The findings suggest that the educational value of GenAI depends less on its mere presence in students’ work than on the way it is integrated into the learning process and the extent to which learners retain responsibility for understanding, evaluating and developing the final outcome. The study therefore highlights the importance of guided, transparent and critically engaged GenAI use in primary education.