.png){kind=icon}
.png){kind=icon}
Return
Active Inference and teacher development
.png){kind=icon}
OA
AI
DOI：10.1080/13664530.2026.2631491.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Teaching unfolds amid uncertainty that cannot be engineered away. Teachers carry internal models of how classrooms work, and when events contradict those models, surprise results. The Active Inference framework, drawn from theoretical neuroscience, formalises this dynamic by treating teachers as agents who generate predictions and learn from the mismatch between expectation and outcome. Surprise is therefore not a planning failure but a signal driving learning and models refinement. The framework also captures something about teacher development: novices cling to routines that minimise immediate surprise, while experts learn to manage expected uncertainty, treating it as informative rather than threatening. This article revisits classic research on teacher thinking to position uncertainty and judgement as central problems, introduces the Active Inference concepts needed to model them, and re-reads three longitudinal case studies to trace the shift from reactive error minimisation to proactive uncertainty management as teachers move from novice to expert.
Keywords:
Active Inference
surprise
uncertainty
free energy
teacher development