Programming MOOCs: it is the nature of the course, not the learner’s profile, that determines the use of generative AI

In two Estonian MOOCs on programming, generative AI is used by learners of all levels, but far more frequently – and for debugging – in the longest and most demanding course.

Marina Lepp interviewed participants from two Estonian-language programming MOOCs: "About Programming", 4 weeks and approximately 26 hours of work (187 respondents), and "Introduction to Programming", 8 weeks and approximately 78 hours (182 respondents). Both courses differ in length, load, complexity of subjects and number of exercises.

General adoption, different uses

The Generative IA, ChatGPT and code-to-lead completion assistants, is widely adopted in both courses, with no significant differences depending on gender, age, educational level or prior programming experience. On the other hand, participants in the longest course make much more use of it, and more use it to debug their code, explain code or find ideas.

The links between AI usage and learning outcomes appear limited in exploratory analyses. The author's conclusion: the context of the course weighs more than the characteristics of the learners, and it is therefore on the pedagogical design and instructions that we must act.