AI systems infer patterns from examples; rule-based systems execute instructions written in advance, and many products combine both.
Learning goals
- Distinguish learned models from fixed rules
- Recognise hybrid systems
- Avoid anthropomorphic explanations
Watch for these misconceptions
- Any automated device uses AI.
- If a system adapts, it must understand.
- AI and robots mean the same thing.
Ask where behaviour came from
Did a programmer specify the decision rule, or was a model fitted using examples? This question is more useful than judging whether the product appears clever.
Expect hybrids
A recommendation may be model-generated but filtered by fixed safety, availability, or business rules. Real systems rarely fit a simple binary.
Use precise language
Models calculate likely outputs; they do not want, know, or understand in the human sense. Metaphors can introduce an idea, but always state where the metaphor stops.
System sorting clinic
- Sort obvious cases first.
- For difficult cases, list the evidence needed to decide.
- Identify the learned and fixed-rule components of hybrid examples.
- Rewrite one anthropomorphic product claim accurately.
Use household examples and focus on ‘follows rules’ or ‘learned from examples’.
Research a product’s documentation and distinguish known facts from inference.
A justified classification that identifies uncertainty rather than guessing.
Check your understanding
Answer each prompt to yourself, then mark whether you can explain it confidently. These are reflective checks, not scored questions.
- What part learns from data?
- What part follows a fixed rule?
- Could the explanation be stated without human-like verbs?
Which metaphors do you use for AI, and what misconception could each create?
Replace ‘the AI knows’ with a more precise description in your next classroom explanation.
Choose a classroom action above, then mark the lesson complete.