Decremental research in the GenAI era: rethinking what learning means when machines do the work / Investigación decremental en la era de la IA generativa: repensando lo que significa el aprendizaje cuando las máquinas hacen el trabajo

Abstract

The rapid adoption of generative artificial intelligence (GenAI) in education has created an urgent need to rethink how we conceptualize and study learning. While GenAI tools can improve academic performance, they also raise a deeper concern: the displacement or erosion of essential cognitive, affective, behavioural and self-regulatory processes. This article introduces the concept of decremental research, a methodological agenda focused not only on what technology enables but also on what it may prevent, bypass or render obsolete in the learning process. In Part 1, we present a conceptual framework grounded in the theory of cognitive offloading, extending it to include affective and behavioural offloading and briefly introducing the concept of decremental research. In Part 2, we use self-regulated learning (SRL) as a case study to illustrate how GenAI may undermine learners’ capacities, and we present three lessons drawn from SRL research. In Part 3, we propose four methodological pillars for investigating what disappears: detailed comparisons with control groups, process data, multimethod and triangulation approaches, and transfer designs. We conclude with a decremental research and design decalogue to help reclaim student agency in the age of GenAI.