AI Practicals, Grade 1 to 12

Grades 1–2 · Smart Machines Around Me

Sort Like a Robot

Aim: Children sort objects by one clear rule, the way a machine follows a rule.
Time: 30–35 min
Materials: Picture cards of fruits, toys and vehicles (printed or drawn), two baskets or chart-paper circles.

Teacher steps

1. Show 10 picture cards. Ask: “How can we group these?”
2. Say: “A robot needs ONE clear rule.” Pick a rule, e.g. “Red things go left.”
3. One child is the “robot” and sorts while the class gives the rule out loud.
4. Give an unclear rule (“nice things go left”). Show the robot gets confused.
5. Let groups make their own rule and test it on another group.
6. Talk: Which rules worked? Why must instructions be clear?
7. Close with one safe-screen habit: ask a grown-up before using any device.

Students learn: Sorting, patterns and clear instructions are how machines work.
Safety: No student names, photos or personal data. Any device or AI tool is used only by the teacher, as a demo on the teacher's own account.
Maps to: No CBSE equivalent (NextBridge foundation unit).

Try it on this website

Grade 3 · Puzzle Detectives

Grid Treasure Hunt

Aim: Write step-by-step moves to reach a target on a grid.
Time: 35–40 min
Materials: A 5×5 grid on the floor (tape or chalk) or on paper, arrow cards (↑ ↓ ← →), a “treasure” object.

Teacher steps

1. Draw the grid and place the treasure.
2. Show arrow cards. Each arrow = move one square.
3. A student writes a sequence of arrows to reach the treasure.
4. A partner acts out the steps exactly. Did they reach it?
5. Add a blocked square. Students fix (“debug”) their steps.
6. Find a pattern: can “→ → →” be written as “3 ×→”?
7. Discuss: computers follow steps exactly, even wrong ones.

Students learn: Sequences, patterns, debugging and grids.
Safety: No student names, photos or personal data. Any device or AI tool is used only by the teacher, as a demo on the teacher's own account.
Maps to: CBSE Computational Thinking, Class 3 (patterns, steps, grids).

Try it on this website

Grade 4 · Mirror, Shape and Step

Mirror Robot

Aim: Use symmetry and multi-step instructions to draw a matching shape.
Time: 35–40 min
Materials: Squared paper, pencils, a small mirror (optional).

Teacher steps

1. Draw half a shape on squared paper with a fold line.
2. Ask students to complete the mirror half. Check with a mirror or fold.
3. In pairs, one student gives step instructions (“2 squares up, 1 right…”) for a shape.
4. The partner draws it without seeing the original.
5. Compare. Where did steps go wrong? Fix them.
6. Sort finished shapes by a rule (symmetric / not symmetric).
7. Discuss: a clear rule plus clear steps = a correct result.

Students learn: Symmetry, sorting rules and multi-step instructions.
Safety: No student names, photos or personal data. Any device or AI tool is used only by the teacher, as a demo on the teacher's own account.
Maps to: CBSE Computational Thinking, Class 4.

Try it on this website

Grade 5 · Think Like a Computer

Teach the Machine (teacher demo)

Aim: See how a computer learns from examples, using Google Teachable Machine.
Time: 35–40 min
Materials: Teacher laptop with webcam and projector, Google Teachable Machine (teachablemachine.withgoogle.com, works in the browser), two objects, e.g. a pen and a bottle.

Teacher steps

1. Unplugged warm-up: show 5 pictures of cats and dogs; ask how students knew the difference.
2. Open Teachable Machine on the teacher's laptop → Image Project.
3. Teacher records ~30 photos of the pen as Class 1 and the bottle as Class 2 (objects only, no faces).
4. Click Train. Explain: the machine is finding patterns.
5. Test with both objects. Then test with a new object. What does it guess?
6. Ask: how could we make it better? (More and varied examples.)
7. Discuss: the machine learned from examples, not from understanding.

Students learn: Machines learn patterns from examples; more varied data gives better guesses.
Safety: No student names, photos or personal data. Any device or AI tool is used only by the teacher, as a demo on the teacher's own account. Do not show any student's face to the camera.
Maps to: CBSE Computational Thinking & AI, Class 5.

Try it on this website

Grade 6 · Meet AI

Doodle Guess with Quick, Draw!

Aim: Find out how AI recognises patterns in drawings, and what kinds of data AI uses.
Time: 35–40 min
Materials: Teacher device and projector, Quick, Draw! (quickdraw.withgoogle.com), Scratch (scratch.mit.edu) for the extension, paper.

Teacher steps

1. Ask: where have you seen AI? List answers on the board.
2. Teacher opens Quick, Draw! on the projector. Students call out how to draw each word; teacher draws.
3. See how fast the AI guesses. Talk about why some guesses fail.
4. Explain data types: drawings, text, numbers, sound, images.
5. Show that many people's doodles trained this AI (the data).
6. Extension: in Scratch, a simple sprite that asks a question and answers using “if”.
7. Close with online safety: never share your name, school or photo with apps.

Students learn: What AI is, types of data, pattern recognition, online safety.
Safety: No student names, photos, voices or personal data in any AI tool. Use the teacher's account and the teacher's screen. Most chatbots need users to be 13+ (some 18+), so students do not sign up; the teacher types the prompts.
Maps to: CBSE Computational Thinking & AI, Class 6.

Try it on this website

Grade 7 · AI Sorts, Guesses and Groups

Sort, Guess and Group with Data

Aim: Try classification, regression and clustering using a small class dataset.
Time: 40 min
Materials: Made-up data cards (e.g. fruits with weight, colour, price), chart paper, optional Google Sheets on the teacher laptop.

Teacher steps

1. Hand out 20 fruit data cards (made-up data, no personal data).
2. Classification: sort fruits into “sweet / sour” using features.
3. Regression: plot weight vs. price on a chart. Draw a line and guess the price of a new fruit.
4. Clustering: without labels, let groups make clusters. Compare groups.
5. Make a bar chart of the clusters (paper or Sheets).
6. Bias check: remove all green fruits from the data. What goes wrong?
7. Discuss: AI is only as fair as its data.

Students learn: Classification, regression, clustering, charts and bias.
Safety: No student names, photos, voices or personal data in any AI tool. Use the teacher's account and the teacher's screen. Most chatbots need users to be 13+ (some 18+), so students do not sign up; the teacher types the prompts.
Maps to: CBSE Computational Thinking & AI, Class 7.

Try it on this website

Grade 8 · My First AI Project

Waste Sorter AI Project

Aim: Follow the AI project cycle to build a no-code model that sorts waste.
Time: 40 min (or two periods)
Materials: Teacher laptop and projector, Google Teachable Machine, sample objects (paper, plastic bottle, peel), project-cycle worksheet.

Teacher steps

1. Problem scoping: what problem? (Waste is mixed in school bins.)
2. Data: plan photo classes — paper, plastic, wet waste.
3. Teacher records photos of each object (no people in frame).
4. Model: train in Teachable Machine.
5. Evaluate: test 10 objects, count right and wrong answers.
6. Fairness: test unusual items (crushed bottle). Who or what does the model fail?
7. Responsible AI: write 3 rules for using this tool fairly and safely.
8. Groups present their project cycle on one page.

Students learn: The AI project cycle, no-code tools, testing, fairness and responsible AI.
Safety: No student names, photos, voices or personal data in any AI tool. Use the teacher's account and the teacher's screen. Most chatbots need users to be 13+ (some 18+), so students do not sign up; the teacher types the prompts.
Maps to: CBSE Computational Thinking & AI, Class 8.

Try it on this website

Grade 9 · AI Builder 1

Data to Chart in Python

Aim: Load a small dataset and find patterns using Python.
Time: 40 min
Materials: Google Colab (colab.research.google.com, needs a Google account; teacher runs it if students are under 13), a small public or made-up CSV (e.g. daily school temperature).

Teacher steps

1. Recap the AI project cycle on the board.
2. Teacher opens Colab and shares the notebook on screen.
3. Load the CSV with pandas: df = pd.read_csv('temps.csv').
4. Find mean, max and min: df.describe().
5. Plot a line chart with matplotlib.
6. Ask: what pattern do you see? What would AI predict next?
7. Extra: ask a generative AI tool (teacher-led) to explain the code; check if it is right.

Students learn: Data handling, basic maths for AI, Python, careful use of generative AI.
Safety: No real personal data in datasets; use public or made-up data. Students follow each tool's age rules; generative AI is used only under teacher guidance.
Maps to: CBSE AI subject 417, Class 9.

Grade 10 · AI Builder 2

Train and Test a Model

Aim: Build a simple classifier, then evaluate it with accuracy and a confusion matrix.
Time: 40 min
Materials: Google Colab, scikit-learn with the built-in Iris dataset.

Teacher steps

1. Explain modelling: the model learns from training data.
2. Load Iris: from sklearn.datasets import load_iris.
3. Split into training and test data (80/20).
4. Train a decision tree model.
5. Find accuracy on the test data.
6. Show a confusion matrix. Which flowers get mixed up?
7. Link to vision and NLP: the same train-test idea works for images and text.
8. Practical support: students write steps in their practical file.

Students learn: Modelling, evaluation, train/test split, accuracy; practical file skills.
Safety: No real personal data in datasets; use public or made-up data. Students follow each tool's age rules; generative AI is used only under teacher guidance.
Maps to: CBSE AI subject 417, Class 10.

Grade 11 · AI Pro 1

Predict and Reflect

Aim: Use machine learning on real data and discuss ethics and careers.
Time: 40 min
Materials: Google Colab, a public dataset (e.g. a simple house-price or rainfall CSV), the NextBridge Career Interest Test (optional).

Teacher steps

1. Load the dataset and clean missing values.
2. Pick one feature and a target. Train a linear regression model.
3. Measure error (mean absolute error).
4. Discuss: who could be harmed if this model is wrong?
5. Write one ethics rule for using this model.
6. Careers link: list 3 AI jobs that use this skill.
7. Start capstone mentoring: each student picks a problem idea.

Students learn: Python, data cleaning, machine learning, ethics, AI careers.
Safety: No real personal data in datasets; use public or made-up data. Students follow each tool's age rules; generative AI is used only under teacher guidance.
Maps to: CBSE AI subject 843, Class 11.

Grade 12 · AI Pro 2

No-Code ML in Orange

Aim: Build a data-mining workflow and tell a story with the results.
Time: 40 min
Materials: Orange Data Mining (orangedatamining.com, open-source desktop app), the built-in Iris or Titanic dataset.

Teacher steps

1. Open Orange. Add File widget and load a built-in dataset.
2. Add Data Table and Distributions to explore the data.
3. Add a Neural Network learner and a Tree learner.
4. Connect Test & Score. Compare which model is better.
5. Add a Confusion Matrix and discuss errors.
6. Data storytelling: make one chart and write 3 lines explaining it.
7. Discuss big data and generative AI: what changes when data is huge?

Students learn: Neural networks, model comparison, data mining, data storytelling.
Safety: No real personal data in datasets; use public or made-up data. Students follow each tool's age rules; generative AI is used only under teacher guidance.
Maps to: CBSE AI subject 843, Class 12.