Hold an item in front of a camera and the matching bin opens. It is an eye-catching demonstration. The more interesting question for a young inventor is what happens when the system gets it wrong.
SUP’s Impact Makers project idea brings together an AI model that identifies types of recyclables and a mechanism that opens the corresponding bin. It gives children a concrete problem to research, build around and explain.
“Sort everything” is too broad a starting point. A learning prototype needs clearly defined categories and examples. Students can discuss which items belong in the demonstration and how the categories relate to the recycling instructions they have researched.
An image prediction is not a complete recycling decision. An object’s appearance does not tell the system every detail about its material or whether it meets local collection requirements. That limitation belongs in the explanation, not hidden behind a successful demonstration.
Imagine a model that recognises an example against a plain table but struggles when the background changes. This is an illustrative test, not a reported SUP result. Students could compare different angles and backgrounds, then test with examples kept separate from training.
The learning question becomes: “What did our examples teach the model, and what have we not tested?” Adding samples is only useful if students also think about their variety and check the result.
The project has another challenge: the correct prediction must lead to the intended bin opening. Students need to consider the model, the instructions and the mechanism together. A confident presentation explains the purpose, demonstrates the prototype and admits what still needs work.
Impact Makers combines AI-supported research, making, testing and pitching. For parents, the value is not that a child has made a machine that never fails. It is seeing them investigate a meaningful problem and support their explanation with a test they actually performed.
Explore the programme: AI FuturePreneur
