ARTIFICIAL INTELLIGENCE TOOLS IN PROGRAMMING EDUCATION FOR PRE-SERVICE COMPUTER SCIENCE TEACHERS: SELECTION CRITERIA AND DIDACTIC SYSTEMATIZATION

Abstract

This article provides a comprehensive analytical review of artificial intelligence tools used in programming education for pre-service computer science teachers and proposes a didactically grounded model for their pedagogical selection and integration. The study addresses a practical and methodological challenge faced by teacher education programs: the transition from occasional and unsystematic use of generative AI services to an intentionally designed instructional environment in which tools are selected according to learning outcomes, stage of student development, and academic integrity requirements. The research synthesizes international policy documents, systematic reviews, and empirical studies on large language models and coding assistants, including ChatGPT, Gemini, Claude, GitHub Copilot, Codeium, Amazon Q Developer, and related instructional applications. Based on comparative analysis and didactic interpretation, the paper introduces a multidimensional systematization of AI tools by educational function, level of intervention, phase of work with code, and type of support. It further substantiates criteria for tool selection in the context of preparing future teachers: didactic relevance, functional fit, stage-appropriate use, verifiability of generated outputs, ethical and legal compatibility, accessibility, and dual professional utility for both personal learning and future classroom practice. The paper also argues that prompt formulation competence should be treated as a core learning outcome in programming methodology courses, because the quality of prompts directly affects explanation depth, student reflection, and the risk profile of tool use. In addition, the article defines organizational and pedagogical conditions required for sustainable implementation, including institutional regulations, teacher professional development, infrastructure support, and monitoring procedures. The proposed model contributes a practical framework for curriculum design and for quality assurance of AI-supported learning activities, and it establishes a methodological basis for future experimental validation in higher pedagogical education. The findings also specify implementation conditions for higher education institutions, including policy alignment, teacher professional development, transparent assessment rubrics, and continuous monitoring of student learning outcomes, which strengthens the practical applicability of the proposed model in curriculum renewal.