A Technology Creates Its Own Pedagogy: Learning as Conversation in the Age of AI
Keynote: II International Meeting of UNESCO Chairs in Open, Digital and Distance Education, "Transforming Higher Education: People-Centred Digital Technologies and Artificial Intelligence." Madrid, 18-19 November, 2026.Every educational technology proposes a pedagogy. Jupyter notebooks, for example, by weaving narrative with code, invited a way of teaching computing as storytelling and live exploration; because they are open source, they also encouraged the sharing of learning resources and became ubiquitous. Generative AI proposes a pedagogy too, and its default is troubling: shaped like conversation, in everyday use it reverts to an antiquated model of instruction, the delivery of answers. Students who use AI as an answer machine develop an illusion of competence: the feeling of understanding without its substance. In this talk, I return to first principles. With Piaget and Papert, learners construct knowledge; with Gordon Pask, understanding is demonstrated by explaining it back in conversation; with Andrea diSessa, literacy is a social infrastructure, not only an individual skill. Read together, they suggest that the question for AI in education is not whether students converse with machines, but who does the explaining. Drawing on my own failed classroom experiment with an AI tutor, and on the innovation that followed—a scalable oral assessment I call the conversational exam, which restores the learner’s voice—I will argue for using AI as an amplifier of human conversation, and for investing in educators as the designers of those conversations.
Authors
- Lorena A. Barba (ORCID: https://orcid.org/0000-0001-5812-2711)
Publication Details
- Journal
- Figshare
- Published
- 2026-11-19
- DOI
- https://doi.org/10.6084/m9.figshare.35063467
- Primary Topic
- Artificial Intelligence in Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00