Prompt-based instructional design in artificial intelligence-supported programming education

Introduction: The rapid integration of generative artificial intelligence (AI) into programming education has transformed prompting from a technical interaction technique into a relevant instructional practice. However, research in higher education remains fragmented. Objective: To synthesize empirical studies on prompt-based instruction in AI-supported programming education, examining conceptualizations, implementation strategies, learning outcomes, and research gaps. Methods: A systematic review following PRISMA 2020 guidelines was conducted. Searches in Scopus and Web of Science identified peer-reviewed publications between 2023 and 2025. Twenty studies met the inclusion criteria and were analyzed through thematic synthesis. Results: Prompting was conceptualized as a technical skill, iterative workflow process, self-regulatory mechanism, computational thinking scaffold, and system-embedded instructional feature. Structured scaffolds, including prompt templates, progressive prompting, and debugging cycles, were associated with improvements in computational thinking and interaction behaviors. Evidence for sustained learning gains remains limited. Conclusion: Prompt-based instructional design is an emerging pedagogical domain whose effectiveness depends on alignment with computational thinking and self-regulated learning principles. Further longitudinal and theory-driven research is needed.