Metasystem Transitions in Education: Human-AI Interactions as Catalysts for Computational Thinking - Figure 2

To complement the qualitative analyses, Figure 2 illustrates the estimated distribution of self-regulation profiles across the two case studies. In the scaffolded condition (Case Study 1), the majority of students were aligned with the moderate self-regulation profile, supported by structured prompts that encouraged productive but not entirely independent use of Gen-AI. A smaller proportion demonstrated high self-regulation, showing more autonomous integration of AI feedback, while only a few fell into the low self-regulation category. In contrast, in the non-scaffolded condition (Case Study 2), the distribution shifted markedly. A larger number of students were classified as low self-regulation, reflecting a tendency to delegate tasks directly to the AI.