Instructional design in teacher education and generative AI: a two-year exploratory study on the alignment of AI-generated learning activities with the Anderson and Krathwohl Taxonomy

The Anderson and Krathwohl Taxonomy (2001) provides a key framework for differentiated instructional design, supporting the development of learning tasks that move beyond the implicit model of the average student. However, limited empirical attention has been devoted to the alignment between generative AI outputs and the cognitive demands defined by the taxonomy. This exploratory study, conducted over the 2024–2026 biennium, examines how 134 pre-service teachers, organised into 32 working groups, evaluated the taxonomic coherence of learning activities generated by ChatGPT, Gemini and Meta AI across the six cognitive process levels. The findings reveal differentiated patterns of alignment across systems, cohorts and cognitive levels. Rather than positioning generative AI as a substitute for teachers’ instructional design competence, the results highlight its potential role as a scaffolding reflective mediator in initial teacher education.