Can AI Be Trusted to Plan Lessons? Auditing the Alignment of Generative-AI Primary English Lesson Plans with China's Curriculum Core Competencies

Generative artificial intelligence (GenAI) is entering primary teachers’ lesson-planning routines, yet to our knowledge no study has audited its output against the competency framework of China’s English Curriculum Standards for Compulsory Education (2022 Edition). Three chatbot models (Claude, DeepSeek, Gemini) each generated lesson plans for four official primary teaching cases from Appendix V of the standards, under a naive and a curriculum-anchored prompt, twice per condition (48 plans; 1,767 activities). Two independent automated coders of different model lineages applied an instrument derived verbatim from the standards (activity-tier kappa = 0.77), and each plan was compared with its official exemplar using Porter’s alignment index. Contrary to earlier audits, the plans were not tier-deficient: transferring-and-creating activities averaged 22.9% of each plan’s coded activities, matching the official cases (23.3%), though naive prompts fell modestly short (17.2%). Alignment gaps concentrated in cultural awareness: under naive prompts only 25% of plans operationalised Chinese-foreign cultural comparison, rising to 96% under curriculum-anchored prompts (cultural-awareness score, p < .001). Document-level competency scores otherwise saturated, and repeated generations varied noticeably. Findings support prompt-anchoring plus teacher review rather than wholesale trust or rejection.

Authors

Institutions

Publication Details

Journal
Journal of Education Humanities and Social Sciences
Published
2026-11-16
DOI
https://doi.org/10.54097/wtwdfd53
Primary Topic
Artificial Intelligence in Education
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Can AI Be Trusted to Plan Lessons? Auditing the Alignment of Generative-AI Primary English Lesson Plans with China's Curriculum Core Competencies

Ziheng Zhang
4 links
Journal of Education Humanities and Social Sciences
Artificial Intelligence in Education
article

Can AI Be Trusted to Plan Lessons? Auditing the Alignment of Generative-AI Primary English Lesson Plans with China's Curriculum Core Competencies

Ziheng Zhang
article en

Abstract

Generative artificial intelligence (GenAI) is entering primary teachers’ lesson-planning routines, yet to our knowledge no study has audited its output against the competency framework of China’s English Curriculum Standards for Compulsory Education (2022 Edition). Three chatbot models (Claude, DeepSeek, Gemini) each generated lesson plans for four official primary teaching cases from Appendix V of the standards, under a naive and a curriculum-anchored prompt, twice per condition (48 plans; 1,767 activities). Two independent automated coders of different model lineages applied an instrument derived verbatim from the standards (activity-tier kappa = 0.77), and each plan was compared with its official exemplar using Porter’s alignment index. Contrary to earlier audits, the plans were not tier-deficient: transferring-and-creating activities averaged 22.9% of each plan’s coded activities, matching the official cases (23.3%), though naive prompts fell modestly short (17.2%). Alignment gaps concentrated in cultural awareness: under naive prompts only 25% of plans operationalised Chinese-foreign cultural comparison, rising to 96% under curriculum-anchored prompts (cultural-awareness score, p < .001). Document-level competency scores otherwise saturated, and repeated generations varied noticeably. Findings support prompt-anchoring plus teacher review rather than wholesale trust or rejection.

Journal of Education Humanities and Social SciencesVol. 65
Shanghai Normal University Tianhua College (CN)
Quality education
Artificial Intelligence in Education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon