The material is not the system
It makes a hypothesis, prediction, evidence and control visible. It never pretends to be AI.
Methodology
Thinking must be visible before it can change.
Sense & Think is a methodology inspired by Montessori education. Much of it begins physically and without screens: a child acts, commits to a prediction and only then lets evidence change the conclusion.
Experience comes first
The principle is not explained in advance. The child experiences it before it is briefly named.
The scaffold must recede
A card, token or bench is temporary. The screen and real system arrive deliberately only when judgement is transferred.
Five verbs, two families
COLLECT, DISTINGUISH and TRACE support investigation. PROTECT and DECIDE lead from knowledge to action.
The map is deliberately small. Children acquire useful questions, not specialist vocabulary.
- COLLECT: What did it have available? What is missing?
- DISTINGUISH: Which test separates the possibilities?
- TRACE: Where did it come from? What supports it?
- PROTECT: What does it really need from me?
- DECIDE: What will I use it for and who is responsible?
Four entries into the methodology
COLLECT works with selection and missing coverage. DISTINGUISH works with the cost of a question and two predictions. TRACE separates a claim, a source and independent support.
For ages 10-12, PROTECT and DECIDE are a reflective map over a concrete task. They do not have a hidden list of single correct answers.
Control is objective only where an objective answer really exists.
Age pathway 6-12
Ages 6-7 form a preparation layer in the target architecture. Its detail remains open. Ages 8-9 are the current core: two hypotheses, a distinguishing test, commitment before evidence and a closed world only.
At ages 9-10 the child names AI, compares it with the simple task and explicitly names the boundary of the analogy. At ages 10-12 a teacher mediates inquiry into real systems. PROTECT, DECIDE, repetition, privacy, consequences and human responsibility are added.
Four pedagogical AI families
Children compare systems that create content, classify, recommend, or predict and assess risk.
This is a pedagogical map, not a complete technical taxonomy. One system can do several things at once.
AI is not a synonym for chatbot. Some systems act on a person without being addressed by them.
Closed and open worlds
In a closed task, A or B is guaranteed to be true. The child can focus on the difference between a weak and distinguishing question.
In real inquiry, there is no guarantee that one of two hypotheses must be true. Where outputs vary, one result is often not enough.
The question transfers. The conclusion from a closed task does not transfer automatically.
Guided discovery
The exact sequence is experience, commitment, evidence, brief naming, contrast and transfer.
The adult does not explain the principle in advance. After the experience, they name it briefly and guide the child to use it on a new surface.
- Experience
- Commitment
- Evidence
- Brief naming
- Contrast
- Transfer
Safety and responsibility
Children do not enter sensitive personal data during activities. Meaning must not depend only on colour, reading, writing or precise motor control.
A child's literacy is not a substitute for safe design. Adults remain responsible for system selection, activity boundaries and decisions with serious consequences.
When working with a live system, an adult operates the interface and checks school and provider rules before use.
The principle remains. The scaffolds recede.
The goal is not to remember a card. The goal is to ask a better question when the card is no longer on the table.