Skip to content
PlayOwl

Computers & Digital Literacy · Grade 10

AI Systems, Data and Limitations

Learners look inside predictive AI systems through small fictional datasets: features and labels, a rule learned from training examples, and why models must be judged on separate test data. They diagnose overfitting, distribution shift and proxy features, discover with simple counts why accuracy misleads when events are rare, and write the limitations a responsible team should document.

  • 4 lessons
  • About 105 minutes
  • AI systems and limitations check: 10 questions
  • Ways to answer: fill in the blanks, choose an answer, trace code, highlight a text, match pairs and type a number
See plans Included with Family, Teacher and School plans. Open in PlayOwl

What learners will discover

  • Identify features and labels in a supervised learning task and check how well a learned rule fits labeled examples.
  • Explain why models are evaluated on separate test data, and distinguish overfitting, underfitting and good generalization from training and test results.
  • Explain how distribution shift and proxy features can make a model fail or treat groups unfairly.
  • Use counts to show why accuracy misleads on imbalanced data, compare with a majority-class baseline, and document a model's limitations.

Lessons

  1. Features, labels and a learned rule 7 steps4 practiceabout 25 min
  2. Training, testing and overfitting 7 steps5 practiceabout 26 min
  3. When the world shifts, and hidden proxies 6 steps4 practiceabout 26 min
  4. The accuracy trap and honest documentation 8 steps5 practiceabout 28 min
  5. AI systems and limitations check 10 questions at the end

Keep exploring

More 10th grade computer skills

Guide for grown-ups