AI Mistakes and Limitations
Machines make mistakes; they only know what they've been shown; bad training data leads to bad results; AI is not magic — just maths on data; showing edge cases and failures
Teaching approaches
Gakuva’s recommended starting approach is selected. You can still compare every option.
1 approach
If an AI incorrectly identified a picture of a muffin as a chihuahua, could your child explain why that kind of mistake happens?
Machines make mistakes; they only know what they've been shown; bad training data leads to bad results; AI is not magic — just maths on data; showing edge cases and failures
Learning resources
Reviewed external links that reinforce this concept. They do not replace Gakuva’s teaching approach.
AI for Oceans
Deliberately probe mistakes: train a limited model, then test odd examples to see confident wrong answers. Discuss that AI can fail without “understanding.”
Proof: Save a screenshot of a wrong or uncertain classification and a parent note of why the learner thinks the model slipped.
Teachable Machine · Image
Adult-guided mistake lab: train with unbalanced or similar-looking classes, then show a misclassification and improve the examples.
Proof: Save a misclassification screenshot and a second screenshot after adding better training examples.
Explore the learning path
1 before · 2 after
This concept
Check understanding
- Give an example of AI making a mistake (voice assistant mishearing, auto-correct error, wrong recommendation)
- Explain that AI mistakes happen because of gaps or errors in training data
- Describe why AI is not magic — it follows mathematical rules applied to data
“If an AI incorrectly identified a picture of a muffin as a chihuahua, could your child explain why that kind of mistake happens?”
Curriculum record
- Type
- Conceptual
- Subject
- Computing
- Domain
- Artificial Intelligence
- Age range
- Ages 7–9