The computational thinking crosswalk

This page exists for one conversation you will have with an administrator, a department head, or a parent: is this a new subject, and where does it fit?

It is not new. The four computational thinking practices that schools and programmes already teach map directly onto working with AI.

The mapping

Decomposition is breaking a problem into solvable parts. Here it is breaking a project into delegable pieces and deciding which parts need a person.

Abstraction is stripping detail to model what matters. Here it is specifying the work: naming constraints, audience, purpose and success criteria while leaving out noise. A good request is an act of abstraction.

Pattern recognition is spotting regularities. Here it is evaluating output, because AI failure modes have shapes. Generic voice, confident errors, surface plausibility, missing context. Learning to see them is quality control.

Algorithmic thinking is designing step-by-step procedures. Here it is designing the workflow itself. Draft, critique, revise, verify is a procedure you author and debug.

Why this matters in practice

One answer is administrative. The other is practical, and it is the one you will actually use.

It answers the placement question. Nobody has to invent a literacy from scratch, and a programme that already teaches CT has the foundation.

It tells you what to do when a learner is stuck. Stuck learners are usually stuck on one of the four. Somebody who cannot get a useful result is almost always failing at abstraction: they have not said what they want. Somebody who accepts bad output is failing at pattern recognition. Naming which one turns a vague struggle into a specific thing to practise.

Where the mapping is imperfect

Say this if someone pushes, because they will be right to.

Traditional CT assumes a system that behaves consistently. A program does the same thing twice. A model does not, and it is fluent when it is wrong, which no compiler is. So pattern recognition carries more weight here than it does in programming, and the patterns are social as much as technical.

The crosswalk earns the word computational. It does not make this identical to teaching programming.

Using it in a session

Do not teach the crosswalk. Learners do not need the vocabulary, and handing it to them turns a practice into terminology.

Use it as your own diagnostic. When you cannot tell why a learner is stuck, ask which of the four they are missing, and you will usually have your answer.