Learning-consequential assessment: making learning count when students use generative AI
Generative AI can enable production without corresponding learning, yet wholesale exclusion from assessed work forgoes opportunities needed for workplace and lifelong learning. This article proposes learning-consequential assessment design. Students may use AI to produce assessed work, but success also depends on showing they possess the understanding their AI-supported work displays. Knowing this may encourage students to use AI for learning through explanation, comparison and challenge, rather than product optimisation alone. The central claim, that the understanding displayed is the learner's own, rests on coherence between their reasons for consequential decisions and their work, with independent performance corroborating where needed. Evidence may also warrant three capability claims: learner-directed production with AI, accountable verification of AI-shaped work and independent capability. Learning Orchestration describes the cognitive-metacognitive, epistemic and ethical judgement through which learners coordinate AI and other resources for learning, production and verification; teacher orchestration describes how teaching and assessment support its development.
Authors
- James M. Wood (ORCID: https://orcid.org/0000-0003-4349-6002)
Publication Details
- Journal
- EdArXiv (OSF Preprints)
- Published
- 2026-11-17
- DOI
- https://doi.org/10.35542/osf.io/jt5uf_v2
- Primary Topic
- Artificial Intelligence in Education
- Type
- preprint