AI Automation ·
How should Korean schools connect optional AI teacher-training modules into a shared lesson design?
A Google News US search for AI education led to UNU Macau’s November 20, 2026 announcement. Teachers may select among three workshops planned for November 28 covering critical thinking, machine-learning practice and lesson design.[1] This is a future programme, not evidence of improved teaching. Our initial question was how choice can coexist with coherence in a school’s shared lesson.
Attending the same programme does not necessarily create shared assumptions
Consider a Korean education support office arranging AI training for teachers across subjects. Optional modules can accommodate needs and available time. But if one teacher selects tool practice and another selects examination of claims, agreement on what would demonstrate student understanding still needs attention. This is a design risk to investigate, not an observed deficiency among participants.
The research question is: when teachers choosing different AI modules jointly design a lesson in a Korean school, where should they establish shared learning goals and the tasks connecting those modules? We compared the overseas choice structure with Korean subject-training arrangements and KERIS’s responsibilities. We surveyed no participants and observed no classes.
The important difference concerned roles, not the number of sessions
The announcement allows separate workshop registration.[1] This shifted our question from whether everyone needs the same lectures to whether each person can explain their contribution to colleagues with different roles. Choice itself has not been shown to be defective. Nor does the announcement establish that compulsory common attendance would automatically create a coherent lesson.
For our proposed image-classification lesson, a teacher working on data should explain which inputs students will change and compare. A colleague should be able to ask how that comparison serves the learning goal. A functioning artefact and a task requiring students to justify a conclusion can differ. This is editorial design analysis, not a student effect demonstrated by UNU.
Korea’s announcement does not give every subject the same starting point
The Education Ministry’s March 16 release describes basic, advanced and specialist subject training for elementary teachers, secondary information teachers and high-school AI-mathematics teachers.[2] Announced operation and measured learning effects are different evidence. It is not a direction to apply one identical course to every subject.
KERIS includes tailored training content, effectiveness evaluation and performance management in its teacher-development work.[3] An institutional responsibility does not establish results for either the overseas programme or our school. It does provide a domestic connection for assigning responsibility for design review and follow-up.
| Design decision | Overseas announcement | Korean official material |
|---|---|---|
| Owner and timing | UNU Macau: announced November 20, planned November 28[1] | Education Ministry subject-training release, March 16[2] |
| Participation structure | Optional registration across three workshops[1] | Specified teaching roles and basic/advanced/specialist stages[2] |
This is not a national ranking of training quality. A Korean organiser should identify the subjects and roles it is supporting before borrowing the overseas choice structure, then allocate a connection activity around the shared lesson. We found no equivalent student-outcome comparison between the programmes.
Option A places a small shared connection task around optional learning
Retain different specialist choices while asking participants to examine the same lesson goal, student task and reasoning evidence. Instead of teaching every tool to everyone, agree what student action will be observed in an example lesson, then bring each module’s contribution back to it. This may suit schools needing cross-subject cooperation with limited common planning time.
The benefit is a way to identify gaps while retaining different expertise. Costs include preparing the shared task, coordination and feedback time. A generic example can miss subject-specific learning, and submission can become paperwork. Requiring every teacher to master all unselected technologies would also erode the benefit of choice.
We propose limiting the shared task to connections between student explanation, inputs/results to compare and the lesson’s response when AI fails. Provide additional foundational support where needed and specify who recognises that preparation as work. We established no universal duration or budget.
Option B retains role-based choices and funds joint design inside the school
Instead of adding common lectures, divide subject, technical-support and assessment-review roles and revise one real unit together. Teachers need not individually complete everything; explaining contributions to colleagues tests the connections. This may fit schools with stable roles and protected collaboration time.
It can also concentrate preparation on one technical teacher and leave the lesson fragile if that person moves. Dividing roles must not divide away accountability. The lesson owner should explain its educational rationale rather than simply repeat the tool specialist’s instructions. School leadership should protect planning time and retain records usable by a replacement colleague.
A connects shared criteria around modules; B commits more resources to unit-level collaboration. The announcement cannot determine which is more effective. A combination is possible, but compare total teacher effort for the same learning goal rather than automatically adding both common work and collaboration.
A pilot separates satisfaction from gaps in the lesson design
We propose that the education support office establish subject needs and prior experience, then choose a representative unit with school lesson owners. Training organisers explain each module’s contribution. Teachers review whether goals, activities, evidence judgments and tool-failure alternatives connect. Determine whether a missing role requires support before transferring the burden to another teacher.
Start with small teams similar in subject, grade and prior experience, working toward the same unit goal. Different preparation time or support can confound a comparison of A and B. First review teacher artefacts. If students are subsequently involved, follow school procedures and data-protection arrangements before bounded classroom observation. Uploading real student material to an external service is not a default pilot requirement.
Proposed indicators include unexplained design assumptions, reasons for revisions following peer review, readiness of a tool-independent alternative and preparation/coordination time by teacher. Claims about student effects require a separate, comparable task in which students explain evidence. Provide support when satisfaction is high but design links are missing; reduce task volume or adjust roles when repeated paperwork grows without a useful connection. These are proposed evaluation criteria, not observed results.
Assigning connection responsibility is better supported than restricting choice
Our conditional answer is to specify the shared lesson goal and module-handoff task rather than require every teacher to attend everything. The overseas programme has not yet run, and the Korean sources are not comparative trials of this choice structure. We cannot conclude that optional modules necessarily fragment learning or that a common course necessarily fixes it.
Actual module combinations, recorded revisions, preparation time and comparable student work could change the judgment. If role-based collaboration establishes equivalent coherence with less burden, reduce common tasks. If a particular combination repeatedly leaves a design gap, strengthen that foundation. Does the school share only a registration list, or responsibility for turning different learning into one lesson?