Program / AI & Smart Machines

AI is powerful, useful, limited, and often wrong.

AI should be tested, questioned, and understood as a tool that produces outputs from patterns. Students need practical experiments and good doubt, not hype.

Ask About Future Programs

Practical questions

What is the machine doing, and how would we know?

Students can compare prompts, inspect patterns, test outputs, collect confident mistakes, and distinguish a useful prediction from understanding.

The central habit is verification: what evidence supports the output, what may be missing, who could be affected, and what a person remains responsible for deciding.

Future topic path

Patterns, predictions, tests, and responsibility.

These are planned public ideas, not claims about an active AI class.

  1. Rules, Patterns, and ModelsFuture foundation
  2. Prediction Is Not CertaintyFuture topic
  3. Prompts as InstructionsFuture topic
  4. Images and ClassificationFuture topic
  5. Games and Bounded DecisionsFuture topic
  6. Data From SensorsActive-program connection
  7. Testing Wrong AnswersActive-program connection
  8. Bias, Privacy, and Human JudgmentFuture topic

Two future previews

Small experiments that expose the limits.

AI & Smart Machines Available later

AI Guessing Game

A future investigation of patterns, predictions, confidence, errors, and fair evaluation.

  • Patterns
  • Prediction
  • Classification
  • Testing
Future Builder Not scheduled
Open future preview
AI & Smart Machines Available later

Chatbot Character Lab

A future critical-writing lab about prompts, character rules, inconsistency, safety boundaries, and verification.

  • Prompts
  • Language
  • State
  • Testing
Future Inventor Not scheduled
Open future preview

Class and Lab direction

Use an output. Then try to break the claim.

A future class would define a bounded question, establish privacy and tool rules, inspect examples, and compare results. Lab work could design harder tests, classify failures, revise a prompt, or connect a limited prediction to a game or robot rule.

Fluent language is not evidence. A generated image is not neutral. A confident answer can still be wrong.

Human judgment remains responsible.

Children can learn about AI safely only when tools, accounts, data, claims, and limits are made explicit. The goal is neither fear nor enthusiasm by default; it is informed use and careful questioning.

Read the AI parent guide

Scratch and Robotics are active first.

Parents may ask about the longer educational direction without assuming AI registration is open.

Contact School of Code