Beyond Understanding: The Under-Explored Role of Affect in Student-AI Interactions

Goda Klumbytė and Claude Draude along with co-authors at the University of Kassel present a small-scale qualitative interview study on students’ use of AI

Human Choice and Computers 2026

IFI Advances in Information and Communication Technology

DOI

Goda Klumbytė Claude Draude

AI Design Interventions for Social Diversity

Education Human-AI interaction Explainability Understandability Affect Human-centred explainable AI

Abstract
This paper presents findings of a small-scale qualitative interview study of students’ use of generative AI tools in their study practices at a German University, and the affective dimensions that emerged in AI understanding and use. We report how affective dimensions manifested through various sites of “(gut) feelings” in AI evaluation and verification as well as in AI use practices, constituting three types of affect-as-resource, namely epistemic resource, access and emotional scaffolding, and affect as interaction resource. We argue that, beyond understanding how individual affective states of the user might affect generative AI understanding and use practices, it is important to critically address affect as a relational domain of agency.