Ask what “learning AI” means
Understanding that a classifier learns from examples is different from implementing and debugging a model pipeline. The first can be explored through discussion, paper activities and suitable supervised tools. The second may require programming, data handling and mathematical ideas. Python is therefore a project prerequisite to discuss, not a universal entry ticket or something a mentor should dismiss automatically. Start with the planned activity. Ask which parts depend on code and which concepts can be explored before the learner is ready to write that code.
Make the starting point observable
A brief can say “can change a short function with help” or “has used blocks but has not written text-based code”. These descriptions are more useful than calling someone a beginner or advanced. Ask the learner to show an ordinary piece of work and describe one decision. Record what help was available. Avoid asking them to perform under pressure for a purchase decision. The purpose is to help the teacher choose the first manageable step, including when that step should be outside the eventual AI application.
Design a core that remains understandable
For a plant-image project, the core could be choosing categories, assembling non-sensitive examples, recording mistakes and explaining why a new photograph is difficult. The mentor might supply a small scaffold while clearly identifying it. A coding-ready learner could then inspect preprocessing or change an evaluation function. Keep extensions separate: collecting more classes, making an interface or deploying publicly should not quietly become conditions of success. The AI4K12 progression charts organise AI ideas by grade bands, but those educational categories do not establish an individual child’s readiness or permission to use a particular product.
Choose teaching that can change pace
Ask the mentor how they will respond if the code becomes a barrier. A good purchasing conversation leaves room to reduce scope, teach a foundation or change tools. It also leaves room for a capable learner to take more responsibility. Agree a short initial pathway and review it against explanation, adaptation and engagement. Age eligibility, supervision and data handling still require their own decisions. A learner can be technically capable and ineligible for a service; another can be eligible but need a different teaching approach. Keeping those distinctions visible helps prevent an impressive output from becoming the only measure of progress.
Sources and limitations
Editorial decision guidance, not an assessment of an individual learner or provider. No lessons were observed. Linked sources provide context, not endorsement of Learning Prospect.
Source review: 18 November 2026. Provider terms can change. Read our methodology.