Checking an AI lesson plan requires an independent check
When teachers revise an AI-generated lesson, are they exercising professional judgment or following the system's lead? A multiple-case study followed five preservice mathematics teachers selected from a larger cohort, using screen records, verbal accounts, interviews and reflections. Across 34 decision episodes, researchers classified 16 as retaining epistemic authority, 12 as sharing it and six as ceding it. Two patterns deserve attention. In four episodes involving two teachers, uncertainty led back to the same AI for verification, creating a closed checking loop. In three episodes involving one teacher, recognition of a problem did not become a concrete instructional response. Editing a lesson therefore did not reliably demonstrate independent evaluation. The study offers a close view of judgment during planning, rather than a prevalence estimate for teachers generally.
Why it matters for leaders
Use professional learning to rehearse two moves: check a doubtful claim against an independent source or worked solution, then write the teaching response it requires. Review those decisions, rather than counting edits as evidence of teacher oversight.
The evidence
Purposeful case selection from 42 preservice teachers enabled detailed reconstruction of decisions. The study connects observable planning activity with teachers' explanations of why they accepted or changed suggestions.
Who was studied
Preservice mathematics teachers in China; 34 analyzed decision episodes.
What the evidence cannot tell us
Five cases in one program cannot establish typical behavior. Recall interviews occurred later, platforms differed, and planning was studied without observing subsequent classroom teaching.
Read the original sources
Summary and leadership analysis from the Week ending 2026-11-06 brief. The edition’s selection notes explain source access and evidence limitations.