Inputs need boundaries
An algorithm should say what kind of information it expects and what happens when something is missing or invalid.
An algorithm should say what kind of information it expects and what happens when something is missing or invalid.
Moving, skipping, or repeating a step can change the output even when all the same words are present.
Normal examples show the basic flow; edge cases reveal conditions the designer forgot to handle.
Start with a familiar non-AI procedure, then add decisions and tests. This makes it easier to distinguish an algorithm from an AI model: every model is used within a larger process, but not every algorithm learns from data.
Which test uncovered the most important hidden assumption?
Do not define an algorithm only as “computer instructions” or imply that every algorithm is AI. Algorithms can be followed by people, and many are explicit procedures that do not learn from examples.
A classroom flowchart is not suitable for decisions about health, identity, school discipline, or safety. Those areas require qualified people, appropriate evidence, and formal safeguards.
Use one related guide, record what the child tested, and return to the Watch & Learn library for the next concept.