<![CDATA[The rapid adoption of generative artificial intelligence in programming education has raised concerns regarding student over-dependence and the erosion of computational thinking skills. This study presents the design and internal validation of SEKAPAI, an AI-based scaffolding platform developed to support computational thinking while promoting responsible use of generative AI. Using an Agile-oriented Research and Development approach, SEKAPAI integrates three adaptive scaffolding modules—Solution Assessment, Code Assessment, and Free Interaction—to deliver context-aware feedback without providing direct solutions. System requirements were derived through stakeholder analysis and translated into a modular, web-based architecture supported by GPT-based services. Internal validation was conducted using comprehensive black-box testing to evaluate functional correctness, feedback behavior, and alignment with computational thinking components. The results indicate that SEKAPAI operates reliably across core system features and consistently implements progressive scaffolding strategies that regulate AI assistance. This study demonstrates how pedagogical scaffolding principles can be operationalized within AI-assisted learning systems and provides a technically feasible reference model for responsible AI integration in programming education.]]>
Copyrights © 2025
