6–8 · Prompting, Models, Responsible Use

How Models Learn

Unit 2 shows how a model actually learns — and where it goes wrong. Across five 30–45 minute lessons, students train models on labeled examples, classify by features, spot bias in unbalanced data, improve a dataset, and defend their data choices.

📘 5 lessons · 30–45 min lessons📄 5 printable activities + certificate

The 5 lessons

  1. 1

    How Models Learn

    30–45 min

    Students explain how a model is trained on labeled examples so it can predict or classify new data.

    Vocabulary: model · training · label · pattern

  2. 2

    Classification Challenges

    30–45 min

    Learners classify items by their features and justify the grouping decisions they make.

    Vocabulary: classification · feature · category

  3. 3

    Bias in Data

    30–45 min

    Kids identify missing or unbalanced data and describe how it leads to unfair or inaccurate AI.

    Vocabulary: bias · representation · unbalanced

  4. 4

    Improving a Dataset

    30–45 min

    Students revise a weak dataset to make it more complete, balanced, and representative.

    Vocabulary: balanced · representative · improve

  5. 5

    Build and Defend

    30–45 min

    Learners present their improved dataset and defend why their choices make for stronger training.

    Vocabulary: justify · evidence · representation

Get the full 6–8 Teacher Pack

Every lesson includes ready-to-teach plans — warm-ups, mini-lessons, activities, assessments, and printables. No prep required.

View school & teacher pricing →← Back to curriculum map