AdaPT: Adaptive Lesson Plan Transformer for Cross-Regional and Differentiated Instruction

Synthesis: Zhang et al. (2026) present AdaPT, an adaptive lesson-plan transformer that adapts lesson plans across regional and differentiated instruction contexts. It improves teacher efficiency while maintaining pedagogical alignment with local curricula — addressing the problem that high-quality lesson plans often mismatch disparate educational contexts, and that current tools generate content from scratch rather than adapting existing plans, increasing teacher workload.

AdaPT uses transformers to adapt lesson plans across regional and differentiated instruction contexts, improving teacher efficiency while maintaining pedagogical alignment with local curricula. Due to educational inequality, high-quality lesson plans often mismatch the needs of disparate educational contexts. Teachers typically modify existing lesson plans to fit new contexts, but current tools instead focus on generating content from scratch, creating additional workload. AdaPT adapts rather than generates from scratch, supporting teacher efficiency and curriculum alignment, connecting to differentiated instruction, equity, and bias considerations in AI-assisted lesson planning.

What this means for practice

• Instructors. Adapt a proven lesson plan to your class instead of prompting for a new one: AdaPT is built on the premise that adapting an existing plan preserves alignment with local curricula, while generating from scratch adds workload, and its teacher participants reported that the system reduced their perceived preparation workload and fit their natural workflow.
• Instructors. Write down what you already know about your students — readiness, interests, and learning gaps — as an explicit profile before transforming a plan, because the formative study found teachers lacked systematic student profile data and relied on personal experience, which made adaptation inconsistent.
• Instructors. Budget for multi-round refinement rather than one-shot acceptance: all six formative-study teachers post-edited AI output, and one generated with one AI and rewrote with another, so build an inspect-and-revise pass into your curriculum planning.
• Designers. Ship multi-level versions of a single plan, a standardized format for quick comprehension, and plug-and-play export — the specific features participants said would lower adoption barriers and counter teacher reluctance.
• Administrators. Back the tool with policy incentives and quality-assurance mechanisms, since participants stressed that technology must be complemented by institutional support to be adopted.

Limitations

• The evaluation is a single study with 12 participants — 9 in-service teachers and 3 education specialists — each completing one realistic lesson plan adaptation task, so no classroom deployment or longer-term use was observed, and the workload benefit rests on teachers' reported perception of reduced preparation workload rather than timer data or independently observed preparation effort.
• Both the formative study (8 participants: 2 education experts, 6 in-service teachers) and the evaluation were run in China and anchored to China's national curriculum standards, which the team pre-populated by decomposing national objectives, so cross-regional transfer was tested inside one national framework only.
• No student-level outcome was measured: judgments of contextual appropriateness come from 3 education specialists, so the system's effect on student learning remains untested.
• The team reports that general-purpose LLMs show limited sensitivity to task difficulty — they produce acceptable tasks but weak gradients between basic and advanced levels — so fine-grained differentiation still needs teacher judgment; the long, detailed prompts that compensate also carry substantial token cost, which the authors name as a practical barrier to adoption.

Citation

Yanjie Zhang, Jiajun Zhu, Minyu Wu, Huamin Qu, Sicheng Song (2026). AdaPT: Adaptive Lesson Plan Transformer for Cross-Regional and Differentiated Instruction.