You unlock a phone with your face. A maps app warns that the road to your nani's house is jammed. A video app somehow knows you will want to watch one more cricket highlight. None of these apps has a person sitting inside it making choices. So who is deciding? Very often, the answer is artificial intelligence, or AI.
In this tutorial you will find out what AI really is, how it is different from ordinary computer programs, and you will play a short game at home that shows the big idea behind it.
So, what is artificial intelligence?
Artificial intelligence is the name for computer programs that do jobs we normally think need human thinking. Recognising a face, understanding spoken words, translating Hindi into English, or guessing what song you might like next are all examples.
"Artificial" means made by people. "Intelligence" here does not mean the computer is wise or has feelings. It means the program can take in information and make a useful guess or decision.
A helpful way to remember it: AI is a computer that has learned patterns from lots of examples, and uses those patterns to make guesses about new things.
Rules versus learning
Most computer programs follow rules that a person wrote, step by step. A calculator is a good example. Someone wrote the exact rule for adding two numbers, and the calculator follows it perfectly every time. It never learns anything new.
Now imagine writing rules to tell whether a photo shows a cat. "Has pointy ears"? So do foxes. "Has whiskers"? So does a rabbit. "Is orange"? Many cats are not. You would need thousands of rules and you would still miss cases.
AI takes a different route. Instead of writing every rule, people show the computer thousands of photos labelled "cat" and "not cat". The program studies them and works out its own patterns. This is called machine learning, and it is the engine behind most of the AI you meet today.
AI around you in India
Look around and you will spot AI in many places:
- Voice assistants that understand you when you ask for the weather in Hindi, Tamil or English.
- Maps apps that guess traffic and suggest a faster route to school.
- Photo apps that group all the pictures of your cousin together.
- Video and music apps that recommend what to watch or hear next.
- Spam filters that move suspicious messages out of your inbox.
Each one learned from huge amounts of examples. None of them "understands" the world the way you do.
What AI is not
AI is powerful, but it has real limits:
- It can be confidently wrong. A chatbot may give an answer that sounds perfect but is simply untrue.
- It only knows what it learned. If it never saw examples of something, it may guess badly.
- It can copy unfairness hidden in its examples. If the examples were one-sided, the AI can be too.
- It does not have feelings, wishes or common sense like a person.
Knowing these limits is just as important as knowing what AI can do.
Try it: play "Teach the Machine"
You need a family member, a pen, and some paper. This takes about 20 minutes.
- Pick a secret rule. Your partner secretly decides a rule for sorting words, for example "words for things you can eat". They do not tell you the rule.
- Collect examples. Your partner writes ten words on slips of paper, each marked YES (fits the rule) or NO. For example: mango YES, chair NO, roti YES, bus NO.
- Study the examples. You are the "machine". Look only at the examples and try to work out the pattern.
- Test yourself. Your partner now shows you five new words without labels. Guess YES or NO for each.
- Keep score. Count how many you got right out of five. That number is your accuracy.
- Make it tricky. Try again with a harder secret rule, like "words with exactly two vowels", and only five examples. Did your score drop?
- Talk about it. When did you make mistakes? Usually it happens when there were too few examples, or when the examples were confusing.
Congratulations: you just did what a machine learning program does. You learned from labelled examples, tested on new ones, and measured how often you were right. Real AI does the same thing, only with millions of examples and a lot of maths.
For parents
This activity needs no screen and no coding, which makes it a good first step for younger children. The key idea to reinforce is that AI learns patterns from examples and can be wrong, so it should be checked, not blindly trusted. Asking "how do you think it learned that?" when your child uses an app is a simple way to keep the conversation going.
What you learned
- AI means computer programs that do tasks we usually link with human thinking, like recognising faces or understanding speech.
- Ordinary programs follow fixed rules written by people. Machine learning programs find patterns from many examples.
- You can find AI in maps, voice assistants, photo apps, spam filters and recommendations.
- AI can be confidently wrong, can copy unfairness from its data, and does not think like a person.
- Accuracy is a way of measuring how often a guess is right on new examples.
Ready for the next step?
This game is the very first idea in learning about AI. In a full course, students go on to see how machines learn in different ways, peek inside a neural network, build small AI programs in Python, and discuss how to use AI fairly and safely. Want to build more? Book a free demo class and see what a live, hands-on AI lesson feels like for your child.