From prompts to programmes: natural language programming as a metacognitive scaffold for GenAI-assisted problem solving

What is it about?

Generative AI tools such as chatbots are increasingly used by students, but they can also encourage overreliance on ready-made answers. This article explores how students can use AI more responsibly and productively through natural language programming: a way of solving problems by clearly describing goals, breaking tasks into steps, checking results and reflecting on decisions throughout the coding process. The study is based on an exploratory pilot in a Master’s Digital Humanities course, where students’ AI interactions were analysed alongside oral project assessments. The observations indicate that when learners are guided to document their reasoning and verification steps, AI can support thinking rather than simply provide shortcuts. The article shows how prompt journals and intermediate artefacts can make students’ reasoning visible and assessable. These ideas are relevant for educators, librarians and institutions designing human-centered digital learning environments that promote agency, accountability and critical AI use.

Featured Image