How Does AI Learn? Patterns, Data, and Smart Guesses

How Does AI Learn? Patterns, Data, and Smart Guesses

Grade 4 · Cybersecurity & AI Education · NYS 4-6.CT.1 / 4-6.IC.1 · 50 Minutes

NYS-Aligned Standards

4-6.CT.1 — Develop a computational model of a system that shows changes in output when there are changes in inputs. 4-6.IC.1 — Describe computing technologies that have changed the world and how they influence cultural practices. NYS Computer Science and Digital Fluency Learning Standards (2020)

Learning Objectives — “I Can” Statements

• I can explain that AI learns by finding patterns in lots of data (examples).
• I can show that when the input changes, the output (the AI’s guess) can change.
• I can give an example of how an AI tool has changed the way people do something.

Essential Question

How can a computer “learn” to make a smart guess, and what happens when the data changes?

Lesson Sequence

Hook / Warm-Up (8 min)

“Guess the Rule” game: teacher sorts objects/words into two groups by a secret rule (e.g., starts with a vowel). Students study the examples (data) and guess the pattern (rule). Connect: this is how AI learns — from examples.

Direct Instruction (12 min)

1. Data = the examples we give the computer. Pattern = what the computer notices. Output = the computer’s smart guess.
2. Example: an AI that tells cats from dogs is trained on thousands of labeled photos. It learns patterns (ears, snout, whiskers).
3. Key idea (input → output): if you change the input (a blurry photo, a new animal), the output guess can change — and can be wrong.

Guided Practice — “Be the Model” (18 min)

Students complete an input → output table for a pretend “Fruit Sorter AI.” Given training examples (round + red → apple; long + yellow → banana), they predict outputs for new inputs and mark which the model might get wrong (round + green = ? ). Students change one input feature and record how the output changes.

Closure (8 min)

Write: “AI learns from ___ . When the input changes, the output ___ . One real AI tool that changed how people do something is ___ .”

SDI & Differentiation Block

Supports for MLLs/ELLs

Entering/Emerging (NYSESLAT Levels 1–2):

• Picture-based input → output table; icons for round/long, red/yellow.
• Sentence frame: “Input: ___ . Output: ___ .”

Transitioning/Expanding (NYSESLAT Levels 3–4):

• Pre-teach: data, pattern, input, output, model, train.
• Sentence frame: “When the input is ___ , the model guesses ___ .”

Supports for Students with IEPs

SDI Adaptation Dimensions: content, methodology, delivery

• Content: Provide a partially completed table with 2 rows modeled.
• Methodology: Use physical sorting cards before the written table.
• Delivery: Read prompts aloud; allow verbal answers; extend time.

Suggested Placement: ICT

Answer Key / Model Responses

Fruit Sorter outputs: round + red → apple; long + yellow → banana; round + green → uncertain (could be apple or lime) — a likely mistake because the training data did not include green fruit.

Closure model: “AI learns from data (examples). When the input changes, the output guess can change. A real AI tool is a translation app that changed how people communicate across languages.”

Alignment Record