Abstract:

Emotion-aware technologies in education have attracted growing research attention in recent years. These systems, designed to detect and respond to students' emotional states, have the potential for supporting personalised learning and improving instructional practices. However, while technical capabilities have advanced rapidly, limited attention has been given to how these systems integrate into teachers’ pedagogical practices or whether teachers can actually understand how these systems reach their conclusions.

This study investigates the role of teachers in emotion-aware educational systems. Emotion-aware educational systems are being developed primarily as technical artefacts, with limited consideration for the pedagogical actors who are expected to use, interpret, and act on their outputs. Therefore, the main objective is to explore strategies for the integration of emotion-aware systems into educational practice by repositioning teachers as central participants in the design and interpretation of affective systems, and by incorporating explainability mechanisms that make system outputs useful for teachers.

This work adopts a conceptual, evidence-informed approach grounded in the research gaps highlighted in our prior systematic literature review on affective computing in education [1]. The review examined 60 studies published between 2018 and 2025 across four major academic databases focusing on how teachers were involved in the design and use of these systems. In addition, the presence of explainability mechanisms was considered as an indicator of whether teachers were genuinely taken into account as end users during system development. Without interpretable outputs for teachers, the practical value of these systems for supporting pedagogical practice remains limited.

The results revealed a consistent pattern of pedagogical disconnection. In more than 80% of the reviewed systems, teachers had no involvement in the operation of the affective system, and fewer than 5% positioned teachers as active users of emotional data for instructional purposes. Further, more than 90% of the studies employed no explainability mechanisms, meaning that the models operated as opaque systems whose outputs could not be easily interpreted or verified by teachers.

Building on these findings, we conceptualize the roles of teachers within these systems. Results present multiple dimensions where the teacher has a central pedagogical role such as interpreter of emotional data (e.g. making sense of classroom affective states), pedagogical decision-maker (e.g. adapting the pace of the class), co-designer of the system (e.g. evaluating the preferred explainability techniques shown by the system) or orchestrator of human–AI interaction (e.g. integrating system’s feedback with other class room signals).

The study proposes strategies to support an active role for teachers in emotion-aware technologies in educational contexts. By conceptualizing key teacher roles and proposing actionable strategies, it addresses the disconnect between technological development and classroom practice, and provides guidance for the design of more transparent, explainable, and pedagogically meaningful emotion-aware systems.

References:
[1] E. Junquera-Talavera, C. Manresa-Yee & J.M. Buades Rubio. Technically Advanced but Pedagogically Misaligned: A Systematic Literature Review of Affective Computing in Education. Artificial Intelligence Review, pp. 1-64 (under review)

Keywords:

Emotion-aware educational system, artificial intelligence in education, explainable AI, affective computing, teacher involvement.