Primary School AI Lesson Examples: Chinese, English, Maths, and General Studies

Three principles for primary school AI teaching

First, a teacher-led interface: younger primary students should not talk to an open-ended chatbot directly. A safer model is for the teacher to operate the AI while the whole class judges the answers, or to use an education tool with a preset framework, so the AI's output passes through the teacher as a layer of checking.

Second, AI gives feedback and students do the thinking: AI is most valuable when it gives instant, individual feedback (comments on writing, hints for solving a problem), not when it produces the answer in the student's place. The lesson design should keep the thinking in the students' hands.

Third, start from an existing topic: do not invent a topic just for AI. Pick a part that is already hard to teach, such as an essay marking cycle that runs too long, or a General Studies lesson where it is hard to extend questioning, and point AI at it, so the results can be measured and recorded.

Chinese example: instant feedback on writing

The old problem in upper primary Chinese writing is the feedback cycle: students get their marking back one or two weeks after handing in, by which time the urge to revise has gone cold. Classroom example: after students finish a first draft, the teacher uses an AI marking tool to produce initial comments in the lesson (on content, structure, and word choice), students revise a paragraph there and then based on the comments, and the teacher moves around the room to guide them. After the lesson the teacher only needs to review the AI comments and the students' revisions, so the focus of marking shifts from circling every word to judging and following up.

Recording tip: keep the set of three, the first draft, the AI comments, and the revised draft. It is the most direct evidence of learning, useful for both parents' day and the report.

English example: tiered reading and speaking practice

Within one class, English reading levels can differ by several year levels. Classroom example: the teacher uses AI to rewrite the same text into three difficulty versions on the spot (different vocabulary loads and sentence lengths, same content), so the whole class reads 'the same story' but each at their own level, and everyone can join the comprehension discussion afterwards.

For speaking, upper primary can use an AI voice tool as an after-class extension of pronunciation practice: students read a set passage aloud, the tool flags the words to watch, and in the next lesson the teacher focuses on the weaknesses the class shares.

Maths example: diagnosing the steps of a solution

What a maths lesson needs to look at is not whether the answer is right, but which step went wrong. Classroom example: when working through a word-problem unit, the teacher gathers students' common wrong solutions and uses AI to generate 'spot the error' exercises, that is, a worked solution with one step off, and students work in groups to find which step is wrong and why. Here AI does the job of quickly generating variant questions close to students' real mistakes, the kind of question that takes the most time to write by hand.

General Studies / Science example: inquiry through fact-checking

General Studies is a natural setting for introducing AI literacy. Classroom example: in an inquiry unit, such as Hong Kong's weather or community facilities, students first observe or measure in the field, then the teacher asks AI the same question, and the whole class compares the difference between 'what AI said' and 'what we measured', discussing why AI is off and how to check it. One lesson meets two goals, scientific inquiry and AI literacy, and can also count towards the student AI literacy activities in the programme.

How to turn these into the teaching demonstrations the programme requires

The circular asks for at least six teaching demonstration examples, covering at least three subjects with two year levels each. The four subject examples above are already a starting base: pick one per subject and run it once at each of two year levels, and the subject and year-level requirements are met. Organise each example in the same format (topic, learning objectives, the role of AI, the lesson flow, samples of student work, teacher reflection), and one example becomes one demonstration you can submit, which can also go straight into an open lesson.