Why language support is the deciding factor

This shows up most clearly in primary and junior secondary. Younger students rely on their mother tongue to think and express themselves, and if a tool only understands English or Putonghua, teachers have to keep translating and rewriting, and it will be shelved within a few lessons.

The cost of a language barrier is not one missing feature; it is the whole tool losing its footing in a local classroom. The more a task needs spoken interaction or Chinese content generated on the spot, the more language support becomes the dividing line.

Three levels to look at

1. Interface language

Whether the interface offers Traditional Chinese, not just Simplified or English. A Simplified interface is just about readable, but its word choices differ from Hong Kong usage, which adds a burden for primary pupils and some teachers, and affects how far parents trust the tool.

2. Speech understanding

Whether the speech feature understands Cantonese, not only Putonghua. For anything involving reading aloud, oral practice or voice commands, this decides outright whether the tool is usable. A tool that only supports Putonghua speech is next to useless in the spoken parts of a Cantonese lesson.

3. Content localisation

Whether the content it produces stays close to the Hong Kong curriculum and its wording. Content built on an overseas or mainland curriculum takes teachers extra time to adapt, and names, examples and marking wording may all need rewriting. Asking the supplier to demonstrate with a unit your school is actually teaching is more reliable than watching an official demo clip.

How a language barrier drags down the three-subject commitment

The programme asks schools to introduce AI-assisted teaching in at least three subjects across two levels each. If a tool's language support only stretches to the English panel, teachers in the other subjects will struggle to pick it up, and the three-subject commitment ends up resting on a handful of people, or cannot be made up at all.

Worse still, a tool that does not run smoothly colours how teachers see the whole programme. If the first teachers to trial it are frustrated by language problems, spreading it to other subject panels afterwards becomes twice as hard. Putting language at the front of the selection is really about protecting the momentum of the rollout.

What to test during a trial

A demo clip tends to use the most ideal examples. What you really need to test by hand during a trial is the following:

  • Feed in your school's real teaching materials and see whether the traditional-character content it produces reads smoothly and whether the wording is local.
  • Read a passage of text or an instruction aloud in Cantonese and see whether the speech feature hears it accurately.
  • Have teachers from different subjects each try it once, rather than leaving the assessment to the English or IT teachers alone.
  • Check that the interface, the documentation and customer support all come in Traditional Chinese, so there is something to fall back on when something goes wrong.

Pin down Simplified versus Traditional and Cantonese versus Putonghua

When a supplier says it 'supports Chinese', press them on whether that is Simplified or Traditional, and Putonghua or Cantonese. These four mean entirely different things in a Hong Kong classroom: traditional characters plus Cantonese is what a local classroom actually runs on, while Simplified plus Putonghua, though still Chinese, may not fit everyday teaching. Ask the question in specific terms, so you do not find the gap only after signing.