From cognitive outsourcing to reallocation: A 3P analysis of student–generative AI engagement in unsupervised assessments

Abstract

The emergence of generative artificial intelligence (GenAI) poses profound challenges to unsupervised assessments in higher education. Yet little is known about how students use GenAI in after-class assessments, and which patterns of use are associated with cognitive disengagement or deeper learning. This study employed Biggs’ presage-process-product  model to examine these processes in authentic assessment contexts. Using a qualitative methodology, we analysed 38 undergraduate students from Japan and China, examining their GenAI usage intentions, dialogue patterns and integration with other learning activities. Our findings show that student engagement with GenAI is not uniform but can be conceptualised as a spectrum bounded by cognitive outsourcing and cognitive reallocation. This spectrum adapts the classic surface and deep approaches to learning in GenAI-supported contexts. We also identified an efficiency paradox as a prevalent misalignment; whereby well-intentioned students default to passive processes and experience cognitive laziness and overreliance. The study operationalised and contextualised the 3P model in GenAI-supported assessment by articulating how presage factors shape distinct process patterns and perceived outcomes. It argues that interventions must move beyond technical skills to foster academic GenAI literacy and dialogue competence to provide explicit guidance that supports a shift from cognitive outsourcing towards more reflective, reallocation-oriented use.

Implications for practice or policy:

• Educators should help students reconceptualise GenAI as a cognitive partner for exploring ideas, rather than just an upgraded search engine for retrieving answers.
• Guidance must move beyond technical skills to explicitly teach the metacognitive strategies and dialogue competence required to manage human-GenAI collaboration.
• Instructors must provide explicit, pedagogical guidance beyond prohibitions, helping students explore effective and ethical GenAI use within specific assessment contexts, shifting them away from outsourcing and towards reallocation.