Danubius International Conferences, 4th International Conference on Education in the Digital Era

Enhancing Mathematics Assessment in Secondary Education through ChatGPT: A Methodological Framework for Digital Item Generation

Gabriela Cristina Branoaea
Last modified: 2025-07-13

Abstract

This paper investigates the integration of ChatGPT, a generative artificial intelligence model, into the process of automated assessment item generation for mathematics education at the lower secondary level. The study aims to enhance the efficiency and pedagogical relevance of formative and summative assessments, in line with the Romanian national curriculum and European digital education strategies. Building on prior research in AI-assisted learning, prompt engineering, and intelligent tutoring systems, this work contributes to the growing body of evidence that supports the use of large language models (LLMs) in educational contexts. Using a design-based research approach, the study combines case analysis and classroom-based experimentation, focusing on how teachers can formulate precise prompts, manage iterative refinement, and validate AI-generated tasks within Moodle and other learning management systems. Findings reveal that ChatGPT can produce curriculum-aligned and cognitively diverse mathematical items, provided that human oversight and instructional expertise guide the process. Key implications include the need for AI literacy and critical digital pedagogy in teacher training. The study offers a methodological framework that institutions can adapt for scalable AI integration. Its originality lies in the practical application of generative AI in mathematics education, offering both immediate utility and long-term value for researchers, educators, and digital learning strategists.