CBSE Class 6 AI Chapter 1: Introduction to AI & Everyday Examples — Complete Teaching Guide

Chapter 1 of the Class 6 AI unit is called Introduction to AI & Everyday Examples, it carries 5 of the 20 introductory AI hours, and the single most important thing about teaching it is this: not one minute of it requires a computer, a coding language, or an AI tool.

This is a chapter about noticing. By the end of five periods a student should be able to walk through a normal day in an Indian town and point at the moments where a machine is making a guess rather than following an order. That is the whole competency. Everything else — the vocabulary, the three types of learning, the human-versus-machine comparison — hangs off that one skill.

The mistake schools make with this chapter is treating it as a technology demo. A projector goes on, ChatGPT is opened, students type a question, everybody claps, and the period ends with a class that thinks AI is a chat box. The syllabus asks for something harder and more durable: conceptual understanding built from examples students already live inside. If you want the full picture of where this chapter sits, read the Class 6 CT & AI syllabus and the wider CBSE AI curriculum for 2026–27.

---

What the syllabus actually says

The CBSE handbook lists the chapter in a single line. Read it exactly as written before you plan anything:

Introduction to AI & Everyday Examples — Understanding what AI is; Difference between AI and Automation; Comparison between human and machine intelligence; Introduction to AI concepts and its types (supervised, unsupervised, and reinforcement learning) — 05 hours.

Four topics, five hours. Notice what is absent: no algorithm names, no maths, no programming, no tool. The word “introduction” is doing real work in that line. Here is what each topic should turn into as a demonstrable outcome.

Syllabus topic	What a student can do by the end
Understanding what AI is	Explain, in their own words, that AI is a machine doing a task that normally needs human thinking, and that it improves by learning from examples rather than being told every rule.
Difference between AI and Automation	Look at any device or app and place it correctly on the automation side or the AI side, and defend the choice with a reason about learning.
Comparison between human and machine intelligence	List at least three things machines do better than people and three things people do better than machines, with an example for each.
Types of AI learning	Match a real situation to supervised, unsupervised or reinforcement learning using the flashcard, button-box and gully-cricket analogies — without any formula.

Only the fourth row usually causes panic in staff rooms. It should not. CBSE is asking for recognition, not derivation. A Class 6 student who can say “that one had labels on the examples, so it is supervised” has met the outcome in full.

---

The concepts, explained

1. What AI actually is

Start with a definition a child can repeat without understanding it, then spend the period making them understand it.

Artificial Intelligence is a machine doing a job that would normally need human thinking — and getting better at that job by learning from examples.

Two halves. The first half is old and easy: calculators do arithmetic, which needs thinking. The second half is what makes it AI. A calculator on the last day of its life is exactly as good as it was on the first. A YouTube recommendation feed on your phone today is different from the one on your cousin’s phone, because it has seen different examples.

Ask the class the question that separates the two: can this machine be wrong in an interesting way? A calculator is never interestingly wrong. Autocorrect changing a Marathi word into something embarrassing is interestingly wrong, because it made a guess. Guessing is the fingerprint of AI.

Write three sentences on the board and have students mark them true or false: “AI is a robot.” “AI is a machine that guesses.” “AI must be connected to the internet.” The answers are false, mostly true, false. That third one matters in schools with poor connectivity — the AI in a phone camera’s face detection works in flight mode.

2. AI versus automation — the distinction everything else rests on

This is the heart of the chapter, and it is where most classes go wrong. Automation means a machine following steps a human wrote down. AI means a machine working out the steps from examples.

Your school is full of automation and almost nobody calls it that. The bell that rings at 10:40 every day is automation. The water pump with a float switch is automation. The photocopier that staples is automation. None of them learn.

Run the two questions in Figure 1 on ten objects from the classroom and the students’ homes. The arguments that break out are the lesson. A traffic signal on a fixed timer is automation; the same junction with cameras that adjust the green time based on how many vehicles are waiting has crossed into AI. Same junction, same pole, different answer.

Be honest with the class about the messy middle. Google Maps contains both: the part that draws the road is automation, the part that predicts the Bengaluru traffic at 6 pm is AI. Most useful products are a sandwich. That admission is not a weakness in the lesson, it is the lesson. This kind of decomposition of a system into parts is exactly the skill built in the Class 6 decomposition module.

3. Human intelligence and machine intelligence

Class 6 students arrive with one of two beliefs, both wrong. Either machines are stupid boxes, or machines are about to become smarter than everyone. The syllabus asks for a comparison, and a comparison means admitting both sides win something.

Make the bars concrete. A machine can check ten crore UPI transactions in a night without getting bored — no human team can. A child who is shown one photograph of a jackfruit will recognise a jackfruit for life; a machine typically needs thousands of pictures. A machine that has seen a lakh of cricket videos still does not know why the crowd goes quiet when a senior player walks out for the last time.

Tell students the sentence to carry away: machines are fast and tireless, people are wise and original, and good work uses both. Do not let the period end on machines-will-take-over. It is not the syllabus and it is not true at this level.

4. The three types of learning — by analogy only

This is the part of the chapter where teachers most often overreach, so let it be said plainly: at Class 6, supervised, unsupervised and reinforcement learning must be taught entirely by analogy. No algorithms. No equations. No accuracy percentages. No Python. The handbook says “introduction to AI concepts and its types” — the word “introduction” rules out the machinery. A student who can sort real-life situations into the three buckets has met the outcome completely.

Use these three analogies and reuse them all year.

Supervised learning — labelled flashcards

You show a child twenty flashcards. Each has a picture and a name written on it: mango, guava, mango, guava. After enough cards, you hold up a new picture with no name and the child says “mango”. The labels were the teaching. That is supervised learning.

Real examples students know: WhatsApp learning to move messages into spam because millions of people marked messages as spam; a bank system learning to spot a fraudulent UPI request because past frauds were labelled as fraud; a hospital tool learning to read an X-ray because doctors labelled thousands of X-rays first.

Unsupervised learning — a mixed box of buttons

Empty a jar of assorted buttons on a desk and say only one thing: “put them into groups.” No names, no correct answer given. Some children group by colour, some by size, some by number of holes. Every grouping is defensible. The machine equivalent finds structure in data nobody labelled.

Real examples: a shopping app noticing that a set of customers all buy the same things around Diwali, without anyone naming that group; a school grouping students by similar performance patterns without deciding the categories in advance. The output is groups, not names. Someone still has to look at the groups and decide what they mean — that is a human job, and it is a good place to point at Chapter 2 on basic data concepts.

Reinforcement learning — gully cricket and the score

Nobody hands you a manual for gully cricket. You play a shot, you get four runs or you get out, and you slowly keep the shots that work. The score is the only teacher. That is reinforcement learning: try, get a reward or a penalty, adjust, repeat.

Real examples: a game where the computer opponent gets harder as you get better; a robot arm learning to pick up a cup after thousands of drops; a delivery app trying different route choices and keeping the ones that finish sooner.

A quick memory hook that works well in class, though CBSE does not prescribe it: labels, groups, score. Supervised has labels. Unsupervised makes groups. Reinforcement follows a score. Three words, one board corner, and most students hold it for the year.

5. Where AI already sits in a student’s day

Before you teach any of this, teach the observation. Students believe AI lives in laboratories abroad. It lives in their pocket and in their family’s kirana payments.

6. AI, machine learning and deep learning

Students will hear these three words used as synonyms on the internet. They are not synonyms. Ten minutes on the nesting saves confusion for the next three years, and it is worth pairing with the definitions in the AI glossary.

---

How to teach it — the five hours

CBSE allots the hours; it does not prescribe how you split them. The distribution below is a recommendation that has worked in classrooms, not a CBSE requirement.

Period	Focus	In-class activity	What students leave with
1	What is AI	AI Spotting Walk — list every machine seen since waking up, then mark which ones guess	A working definition and a personal list of examples
2	AI versus automation	Sorting Table — 15 object cards sorted into two columns with a written reason	The two-question test from Figure 1
3	Human versus machine intelligence	Beat the Machine — timed arithmetic race, then a common-sense riddle race	Three human strengths and three machine strengths, with examples
4	The three types of learning	Flashcards, button box, gully cricket score — all three analogies run live	Labels, groups, score — and the ability to sort situations
5	Consolidation and assessment	Group presentation: “AI in my day” poster, plus the worksheet	A finished artefact and a completed worksheet

Period 1. Do not define AI in the first five minutes. Ask students to list every machine they have used since waking. You will get twenty to forty items in three minutes. Then draw one column for “does exactly what it is told” and one for “seems to guess”. The definition lands after the sorting, not before.

Period 2. This period is the spine of the chapter. Prepare fifteen slips: ceiling fan regulator, washing machine, calculator, ATM, IRCTC seat printing, school bell, YouTube feed, Google Maps traffic estimate, autocorrect, UPI fraud alert, DRS ball tracking, WhatsApp spam filter, water pump float switch, photocopier, monsoon forecast. Groups sort them and must write a one-line reason. Collect the disagreements and settle them on the board.

Period 3. Run two races. First, twenty three-digit multiplications against a calculator — the class loses badly. Second, a riddle a calculator cannot touch: “Amma left the dosa on the stove and went to answer the door. What should she worry about?” The class wins instantly. Then draw Figure 3 on the board from memory.

Period 4. Physical materials only. Flashcards for supervised, a jar of buttons for unsupervised, and a five-minute reenactment of learning a shot from the score for reinforcement. Say the boundary out loud to the class: we are learning what these three words mean, not how they work inside. That comes years later.

Period 5. Groups build a chart-paper poster of a day in their life, marking every AI touchpoint and labelling each as supervised, unsupervised or reinforcement where they can. This doubles as the assessment artefact and as the seed for the interdisciplinary project hours described in the CT and AI pillar guide.

---

Unplugged activities

Activity 1 — The Flashcard Machine

Time: 20 minutes. Materials: 20 index cards or torn notebook pages, a pen.

1. Draw or write a fruit on each card — ten mangoes, ten guavas, drawn differently each time. Write the name on the back of fifteen cards. Leave five with no name.
2. Pick one student to be the machine. They sit facing the board and may not speak.
3. Show the machine the fifteen labelled cards one at a time, saying the name each time.
4. Now show the five unlabelled cards. The machine writes its guess for each.
5. Score the guesses. Repeat with a second student who is only shown five labelled cards first.

Discussion question: the second machine saw fewer examples and made more mistakes. What does that tell you about how much data an AI needs, and about who is responsible if the labels were wrong?

Activity 2 — The Button Box

Time: 20 minutes. Materials: a jar of mixed buttons, or bottle caps, or torn paper of different shapes. Free.

1. Tip the buttons onto four desks, roughly forty per group.
2. Give exactly one instruction: “Put these into groups.” Refuse to answer any question about how many groups or what kind.
3. Give five minutes. Then each group explains its grouping rule.
4. Write every group’s rule on the board. Count how many different rules the class produced.

Discussion question: no group was told the right answer, and every group produced something sensible. If a machine grouped your school’s students this way without being told the categories, who decides what those groups mean — and what could go wrong?

Activity 3 — Gully Cricket Scoreboard

Time: 25 minutes, best in the ground or corridor. Materials: a soft ball, a bat or a plank, chalk. Free.

1. Mark three zones on the ground: 1 run, 4 runs, out.
2. A student bats with eyes on the zones and no coaching whatsoever. Ten balls.
3. A partner records the score after every ball on a chalk table.
4. Bat a second set of ten balls. Compare the two totals.
5. Ask the batter to describe, in one sentence, what they changed and why.

Discussion question: nobody taught the batter a technique — the scoreboard did. Name one thing you have learnt this way, and one thing that would be dangerous to learn this way. The second half of that question links directly to Chapter 4 on AI ethics and digital responsibility.

---

Worksheet

Part A — Recall (1 mark each)

1. Complete the sentence: AI is a machine doing a job that normally needs ______ thinking, and it improves by learning from ______.
2. Name the type of learning that uses labelled examples.
3. Name the type of learning that groups things without being given any names.
4. Name the type of learning that improves by following a score or reward.
5. State one thing machines do far better than people.
6. State one thing people do far better than machines.
7. True or false: every AI system must be connected to the internet.
8. Which is bigger, machine learning or artificial intelligence?

Part B — Apply (2 marks each)

9. For each, write AI or automation with a one-line reason: (a) a school bell on a timer, (b) autocorrect in Marathi, (c) a photocopier, (d) a WhatsApp spam filter.
10. Google Maps says the school bus will take 42 minutes today and took 31 minutes yesterday on the same road. Explain why this is AI and not automation.
11. A shopping app groups customers who all buy raincoats and umbrellas in June, without anyone naming that group. Which type of learning is this? Give your reason.
12. A video game opponent becomes harder each time you beat it. Which type of learning is this? Give your reason.

Part C — Think (3 marks each)

13. A friend says “AI is a robot.” Write three sentences to correct them, using one example from your own home.
14. An AI is trained to recognise Indian fruits using only photographs taken in Kerala. It is then used in Kashmir. What might go wrong, and whose responsibility is it?
15. Describe one task in your school that is automation today and could be improved by AI. Say honestly what could go wrong if it were.

Answer key

1. human; examples (or data)
2. Supervised learning
3. Unsupervised learning
4. Reinforcement learning
5. Any of: speed of calculation, working without tiredness, checking huge numbers of records
6. Any of: common sense, creativity, understanding context, learning from a single example
7. False. Face detection in a phone camera works offline.
8. Artificial intelligence. Machine learning sits inside it.
9. (a) automation, fixed time, never changes. (b) AI, it learns from what you and others type. (c) automation, same steps each copy. (d) AI, it learns from messages people mark as spam.
10. The estimate changes because the system learns from live and past traffic data. Automation would give the same number every day.
11. Unsupervised. Nobody gave the group a name or a label in advance; the system found the pattern itself.
12. Reinforcement. The system adjusts based on the outcome of each game, which acts as its score.
13. Accept any answer that says AI is software not a body, that a robot is a machine that may or may not contain AI, and that gives a valid home example such as autocorrect or a recommendation feed.
14. It may fail on fruits, lighting and varieties it never saw. Responsibility sits with the people who chose the training data, not with the machine.
15. Accept any reasoned answer. Award the third mark only where a genuine risk is named, such as a wrong prediction affecting a student unfairly.

---

Common misconceptions

What students or teachers believe	What is actually true	How to correct it
AI is a robot	AI is software. A robot is a body. Most robots on a factory floor contain no AI, and most AI has no body at all.	Point at a UPI fraud alert and a washing machine in the same minute. One has AI and no body, one has a body and no AI.
AI thinks like a human	AI finds patterns in numbers. It has no understanding, no intention and no idea what any of it means.	Ask what autocorrect thought your Hindi word meant. It thought nothing. It counted which word usually follows.
ChatGPT is all of AI	Chat tools are one small, recent branch. Spam filters, weather models, DRS tracking and recommendation feeds are all AI and none of them chat.	Run Figure 5. Only one of the seven daily touchpoints is a chatbot.
Automation equals AI	Automation follows written rules. AI works rules out from examples. Automation never improves.	Use the two-question test in Figure 1 on the school bell and the YouTube feed.
AI is always right	AI produces a prediction, and predictions can be confidently wrong, especially on data unlike the training data.	Show a genuine autocorrect failure in Hindi or Marathi. The class will supply plenty of their own.
AI learns on its own, from nothing	Someone collected the data, someone chose the labels, someone decided what counts as a good outcome. People are inside every step.	After the flashcard activity, ask who wrote the names on the cards, and what would happen if they wrote them wrongly.
Teaching this chapter needs computers	Every outcome in Chapter 1 can be met with chalk, paper and buttons. No tool is required at Class 6.	Run all three unplugged activities in a room with no electricity and check the worksheet scores.
My child needs Python for this	There is no coding in the Class 6 AI unit at all. Coding is not what is being assessed.	Show parents the syllabus line and the learning-outcome table above at the first parent meeting.

---

Assessment

Assess three things: whether the student can classify examples correctly, whether the reason behind the classification is sound, and whether they can produce their own examples rather than repeating yours. The third one is the real test of understanding.

Do not assess terminology recall alone. A student who writes “reinforcement learning” perfectly but cannot say why the gully cricket score is the teacher has not met the outcome. This stage is competency-based and largely internal — schools judge it through classwork, activities and portfolio evidence rather than a formal external paper.

Criterion	Beginning	Developing	Secure
Explaining what AI is	Repeats a definition without examples, or says AI is a robot	Gives a correct definition with one example supplied in class	Defines AI in own words with an example from their own life not used in class
Separating AI from automation	Sorts by guesswork; reasons are missing or wrong	Sorts most items correctly, reasons are thin	Sorts correctly including mixed cases, and names the learning as the deciding factor
Human and machine strengths	Says one is simply better than the other	Names two strengths on each side	Names three on each side with an example, and explains why both matter together
The three types of learning	Confuses the three or recalls only the names	Matches given situations to the right type with prompting	Matches new situations independently and explains using labels, groups or score

If you want to check whether your school is ready to run this kind of internal, competency-based assessment at all, the school AI readiness score is a fifteen-minute audit.

---

Interdisciplinary links

Class 6 carries 40 hours of interdisciplinary projects. Chapter 1 is the natural feeder for them, because “spot the AI, then argue about it” transfers into every subject. These are project ideas, not CBSE-prescribed projects.

• Mathematics. Collect the class’s daily screen time for a week, group the data, and present it as a chart. Then ask which grouping the class chose and whether a machine would have chosen the same one — a direct bridge to the unsupervised analogy and to Chapter 3 on pattern recognition and decision making.
• Science. Compare the local monsoon forecast against what actually happened, for fourteen days. Students record predicted versus actual rainfall and calculate how often the model was right. This teaches prediction and error better than any lecture.
• Social Science. Map the school bus routes and interview the driver about which roads jam and when. Compare the driver’s knowledge with what a maps app predicts. Which one knows about the Tuesday market?
• Languages. Type ten Hindi or Marathi sentences with autocorrect on and record every mistake it makes. Classify the mistakes. Write a short piece arguing whether autocorrect helps or harms a regional language.
• Art and design. Build the “AI in my day” poster from Period 5 as a proper design brief — legible from three metres, colour-coded with text labels, one figure per touchpoint. Figure 5 is the model.

All five draw on the same four computational thinking skills the other 40 hours cover: pattern recognition, abstract thinking, algorithmic thinking and decomposition.

---

Frequently asked questions

Do we need computers or an internet connection to teach Chapter 1?

No. Every learning outcome in this chapter can be met with chalk, paper, flashcards and a jar of buttons. A demonstration is optional and is not required by the syllabus at Class 6.

Is there any coding or Python in Class 6 AI?

There is no coding in the Class 6 AI unit. The content is conceptual, applied and ethical. Tell parents this clearly at the first meeting, because coaching centres will suggest otherwise.

How deep should supervised, unsupervised and reinforcement learning go?

Analogy depth only. Labelled flashcards, a mixed box of buttons, and learning gully cricket from the score. No algorithms, no mathematics, no accuracy figures. A student who can classify a new situation correctly has met the outcome.

How many hours does this chapter get, and where does it sit?

Five hours. It is the first of four chapters in the 20-hour introductory AI component, which sits alongside 40 hours of advanced computational thinking and 40 hours of interdisciplinary projects in the 100-hour Class 6 allocation.

What is the single hardest idea for students in this chapter?

That the same device can contain both automation and AI. Google Maps drawing the road is automation; Google Maps predicting the travel time is AI. Teach the mixed case deliberately rather than avoiding it.

How is this chapter assessed?

Through classwork, activities and portfolio evidence. This stage is competency-based and largely internal. Assess whether a student can classify examples, justify the classification, and produce fresh examples of their own.

---

PiyushAI builds classroom-ready material for CBSE’s Computational Thinking & AI rollout. Start with the CT & AI pillar guide, then move to teacher training for CBSE, AI literacy for teachers and AI literacy for students.

Related: Chapter 2: Basic Data Concepts · Chapter 3: Pattern Recognition & Decision Making · Chapter 4: AI Ethics & Digital Responsibility · Class 5 syllabus · Class 7 syllabus

Source: CBSE Computational Thinking & Artificial Intelligence handbook, Class 6 (2026–27), cbseacademic.nic.in.