Abstract:

The rapid evolution of artificial intelligence (AI) tools is transforming how software is conceptualized and built. "Vibe coding”, a colloquial term for AI-assisted development where natural language prompts guide code generation enables educators and learners to rapidly prototype ideas and reach Proof of Concept stages without deep syntactic expertise.

This shift challenges traditional computer science pedagogy, which has long prioritized manual coding proficiency. At the Faculty of Science, University of Split, we conducted a comparative study assessing understanding of basic programming concepts among future educators. Results indicate that while AI tools significantly lower barriers to functional output, learners who focused exclusively on syntax; making the code run; struggled with problem decomposition, requirement specification, and architectural reasoning. Conversely, participants focused on system analysis; defining inputs, outputs, constraints, and user flows; used AI more effectively to iterate, debug, and scale solutions. These findings suggest an urgent need to reframe programming education for the next generation of educators: from teaching how to code to teaching how to think like a system designer.

This paper proposes a pedagogical framework for integrating "vibe coding" into teacher training curricula, emphasizing competencies such as problem framing, prompt engineering, validation logic, and human oversight of AI-generated code. Grounded in emerging machine teaching principles, the framework positions the educator as a designer of learning interactions between humans and AI systems. As a result we propose practical strategies for curriculum redesign, assessment rubrics focused on analytical rather than syntactic mastery, and collaborative classroom models where AI acts as a co-worker rather than a replacement for critical thinking. We argue that empowering educators with system analysis skills, not just coding fluency, better prepares them and their future students for a technology-rich, AI-augmented world. The study concludes with recommendations for policy, faculty development, and cross-disciplinary collaboration to ensure that programming education evolves in step with the tools reshaping our digital landscape.

Keywords:

Artificial Intelligence, Computer Science Education, Teacher Training, System Analysis, Machine Teaching.