DRIVE — TU Eindhoven's Framework for Assessing Learning Through Students' Interactions with Generative AI

DRIVE (Directive Reasoning Interaction and Visible Expertise) is a conceptual framework and taxonomy developed at Eindhoven University of Technology to let educators judge domain learning from how students steer and use generative AI chatbots, rather than only from the final text.
The authors ran a multi-methods analysis of 1,450 annotated GenAI interactions drawn from 70 graded essays in STEM writing-intensive courses. They report a strong positive correlation between the quality of GenAI interactions and final essay scores, supporting the feasibility of the approach.
The taxonomy surfaced distinct usage profiles: a 'targeted improvement partnership' (text refinement) linked to high essay scores, a 'collaborative intellectual partnership' (idea development) linked to high interaction scores, and 'basic information retrieval' or 'passive task delegation' linked to below-average scores. The authors note that output-focused assessment can reinforce text optimisation, while process-focused assessment may reward exploratory partnership with AI.
The study was published in Computers and Education: Artificial Intelligence (Vol. 9, December 2025). Caveats: the sample is small (70 essays), the evidence is correlational and from one institution's STEM courses, so transfer to other disciplines and levels is untested.

Read the full analysis: https://research.tue.nl/en/publications/assessing-students-drive-a-framework-to-evaluate-learning-through/