Educational Psychology

Unpacking the “Black Box” of AI-Assisted Feedback: Teachers’ Feedback Perceptions, Emotions, Actions, and Literacies

Artificial intelligence is increasingly embedded in educational feedback practices, offering new possibilities for timely, scalable, and personalised responses. These developments extend a long-standing shift in feedback research from transmission models toward understanding feedback as a process that supports learning (Boud & Molloy, 2013; Hattie & Timperley, 2007). Within this shift, learner-centred perspectives and the concept of student feedback literacy have provided important advances in explaining how learners engage with and use feedback (Carless & Boud, 2018; Molloy et al., 2020).

However, recent developments suggest that this account remains incomplete without a more explicit focus on teachers. Research on teacher feedback literacy indicates that teachers play a central role in shaping feedback processes through design, interaction, and judgement, yet this line of work remains conceptually diverse and unevenly developed across contexts (Boud & Dawson, 2023; Carless & Winstone, 2023). Empirical studies tend to focus on local practices such as written comments or feedback talk, often in small-scale or discipline-specific settings (Heron et al., 2023; Lee, 2021), with limited integration across broader feedback systems or technological environments. As a result, there is still limited understanding of how teachers design and sustain feedback processes in ways that enable student uptake across contexts.

At the same time, feedback research has increasingly highlighted the importance of emotional, relational, and motivational dimensions. Feedback is interpreted through learners’ sense-making processes, shaped by trust, perceived support, and beliefs about improvement (Fong & Schallert, 2023). These processes are not independent of teaching practice. Rather, they are influenced by how teachers frame feedback, respond to learners, and create conditions for engagement (Xu & Carless, 2017; Payne et al., 2023). This suggests that feedback cannot be fully understood without examining how teachers enact these relational and pedagogical dimensions in practice.

The emergence of artificial intelligence introduces new complexity into this landscape. AI systems can generate feedback at scale and with high levels of responsiveness, but their educational value depends on how they are interpreted, mediated, and integrated into teaching. Evidence indicates that teachers’ trust, self-efficacy, and understanding play a critical role in shaping their perceptions and use of AI technologies (Viberg et al., 2025). Without sufficient confidence and critical awareness, there is a risk that AI-generated feedback may be either over-relied upon or underutilised, limiting its pedagogical value.

From a broader perspective, global policy work underscores the urgency of redefining teachers’ roles in relation to AI. The introduction of AI is reshaping educational interactions into a more complex teacher–AI–student dynamic, requiring teachers to exercise judgement, maintain human agency, and ensure ethical and responsible use of technology (UNESCO, 2024). This includes the capacity to evaluate AI outputs, guide students’ engagement, and design learning environments that preserve meaningful human interaction. Yet, there remains limited empirical research that examines how teachers enact these responsibilities in feedback contexts.

Taken together, these strands of research point to a clear need to move beyond isolated perspectives. Feedback needs to be understood as a process that is simultaneously cognitive, emotional, relational, and increasingly technological. Teachers are central to this process, not only as providers of feedback, but as designers of learning environments and mediators of how feedback is interpreted and used.

This Special Issue addresses this need by focusing on teachers’ feedback perceptions, emotions, actions, and literacies in AI-assisted environments. It aims to advance a more comprehensive understanding of feedback as a process shaped through the interaction between teachers, students, and emerging technologies. By integrating perspectives from educational psychology, assessment, and learning sciences, the Special Issue seeks to strengthen conceptual clarity, support methodological development, and provide insights that are relevant to both research and practice.

We welcome submissions on topics including, but not limited to:

• Teachers’ perceptions, beliefs, and trust in AI-assisted feedback
• Teachers’ emotional experiences and their role in shaping feedback engagement
• Teacher feedback literacy in technology-mediated contexts: conceptualisation and measurement
• Design and orchestration of feedback processes involving AI, teachers, and peers
• Interplay between teacher and student feedback literacy
• Teachers’ roles in supporting students’ interpretation and use of feedback
• Ethical and pedagogical considerations in AI-assisted feedback
• Institutional and disciplinary influences on feedback practices
• Longitudinal and intervention studies on teacher feedback literacy
• Comparative studies of AI-assisted and human feedback
• Systematic reviews and meta-analyses focusing on teacher-related dimensions of feedback
• Engagement with Emerging Conceptual Frameworks

Recent conceptual developments, such as the TEP-AIED framework proposed by Hwang et al. (2026) and the AI-assisted feedback mechanisms model by Ba et al. (2025), offer valuable perspectives for examining methodological and pedagogical issues in AI-assisted feedback. This Special Issue welcomes studies that draw on, critically examine, extend, or empirically test these and other relevant frameworks to advance understanding of the “black box” of AI-assisted feedback across diverse educational contexts.

• Cross-Cultural Validation and Adaptation of Feedback Instruments

This Special Issue also welcomes studies examining the applicability, validation, adaptation, or cross-cultural use of feedback-related instruments including, but not limited to, the Teacher Feedback Literacy (TFL) Scale (Istencioglu et al., 2026), in AI-assisted feedback contexts across diverse educational, disciplinary, and cultural settings.