AI Open Lesson Checklist: 12 Things to Do

The open lesson is the most public of the five deliverables under the AI for Empowering Learning and Teaching Funding Programme, and the hardest to patch up at the last minute. At least three of them, watched by colleagues, with results on show in the room. How well an open lesson goes depends mostly on the preparation beforehand. The checklist below is ready for a subject panel to use as it stands.

The role of the AI open lesson in the programme

Under Circular Memorandum 221/2025, participating schools must hold at least three open or demonstration lessons, open to colleagues and to other schools. Its function works on two levels. Inside the school, an open lesson forces a subject panel to refine AI use from 'one teacher's habit' into 'a teaching design that can be shown'. Outside, it is one of the most convincing pieces of evidence in the interim and final reports, because it has a date, observers, and a record of the lesson.

The three open lessons are best spread across different school years and different subjects, rather than crammed into three sittings of the same subject. That fits the 'at least three subjects' commitment and gives each lesson enough time to prepare.

Eight things to do before the lesson

1. Pick a lesson where AI genuinely improves learning: AI should solve a teaching difficulty the topic already has, such as instant feedback or tiered practice, rather than an activity slotted in to show off a tool.
2. Write clear learning objectives: an observer should see at a glance what students learned. If all they can see is what students used, the teaching value of the lesson cannot be shown.
3. Run the whole lesson through at least once with another class in the same year level, noting how long the AI part actually takes.
4. Rehearse for technical failure: a dropped connection, a login that fails, unsuitable AI output. Each needs a backup plan, and an offline worksheet is the simplest insurance.
5. Check student accounts and data settings: in an observed lesson especially, make sure no student's personal data is put on display.
6. Prepare an observation sheet: turn 'how AI supports learning' into three or four specific observation questions to steer the post-lesson discussion.
7. Notify and invite: beyond your own colleagues, you can invite a partner school to observe. With a partner school present, the lesson carries more weight in the report.
8. Arrange to record the lesson: photos or video (following the school's established consent procedure), how student work will be collected, and who is responsible for the discussion notes.

Two things on the day

9. Spend the opening minute telling observers the learning objectives, where AI appears in the lesson, and why it is used, so they hold the right expectations.
10. Teach at the pace of your trial run, and if technology fails, switch straight to the backup without dwelling on it. Handling a fault calmly is a good demonstration in itself.

Two things after the lesson

11. Hold the post-lesson discussion the same day: talk through the questions on the observation sheet and record what you would change next time. Those notes are direct material for the report.
12. File everything within seven days: the lesson plan, teaching materials, samples of student work, the observation sheet, the discussion notes, and the photos, all placed in the school's evidence folder, so you are not rummaging through boxes when the report is due two years later.

The four most common mistakes in an AI open lesson

The first is the 'tool exhibition lesson': three or four AI tools packed into one lesson, with students busy switching platforms while the learning objective blurs. A good AI open lesson usually uses one tool to solve one teaching difficulty.

The second is having no backup plan: a campus network is most likely to fail when dozens of people connect at once, and an open lesson is exactly that kind of occasion.

The third is skipping the discussion: the lesson ends, people leave, and with no discussion notes the report loses its evidence of professional exchange, while the promise of three sharing sessions loses ready-made material.

The fourth is preparing too late: the teacher meets the tool for the first time two weeks before the lesson. An open lesson should be led by a teacher who has used the tool day to day for more than a term, because fluency cannot be built in a short time.

AI Open Lesson Checklist: 12 Things to Do Before and After the Demonstration