What is AI, simply?

A simple definition

Artificial intelligence, or AI, is a name for computer programs that can do jobs we used to think needed a human brain. Recognising a face in a photo, turning your voice into text, suggesting the next word as you type, or answering a question in full sentences are all examples.

The important word is learns. A normal program follows instructions that a person wrote out step by step, like a recipe. An AI program is instead shown a huge number of examples and works out the patterns for itself.

Tip: When you hear "AI", think "a very fast pattern-spotter", not "a robot that thinks". You will rarely be far wrong.

Learning from examples, like a child

Think about how a small child learns what a dog is. Nobody gives them a rulebook that says "four legs, fur, a tail". They simply see lots of dogs, and lots of things that are not dogs, and slowly their brain picks up the pattern. Eventually they can point at a breed they have never seen before and say "dog".

AI is trained in a similar way. To build a program that spots cats in photos, engineers feed it many thousands of pictures labelled "cat" or "not cat". Each time it guesses wrong, it adjusts itself a tiny bit. After millions of small adjustments it becomes good at the task. This process is called training, and the finished result is called a model: a large file full of numbers that captures what was learned.

Where you already meet AI

You probably use AI many times a day without noticing:

• Your phone camera finds faces and brightens them, and your photo app can search for "beach" or "birthday".
• Spam filters in your email learn what junk looks like and move it out of your inbox.
• Maps predict traffic and suggest a faster route.
• Streaming and shopping sites recommend films or products based on what people like you chose.
• Voice assistants turn your speech into words and work out what you are asking for.
• Chatbots such as AI assistants write replies, summaries and emails. See how chatbots work.

What AI is good at, and what it is not

AI shines at tasks with lots of examples and clear patterns. It can read a thousand pages in seconds, sort photos, translate between languages, and draft text quickly.

But it has real limits that are worth knowing:

• It does not understand. It predicts what is likely, based on patterns. That is different from knowing what is true.
• It can be confidently wrong. A chatbot may invent a fact, a date or a book title and present it as certain. People call this "making things up" or hallucinating.
• It copies the habits of its examples. If the training data was unfair or one-sided, the AI can repeat that unfairness.
• It has no common sense or feelings. It cannot care about you, even if its words sound warm.

Tip: Treat AI like a quick, eager helper who has read a lot but sometimes mixes things up. Useful, but check anything important.

Common myths

"AI is alive or conscious." No. Today's AI is software running on computers. It has no wishes, no awareness and no inner life.

"AI knows everything." No. It only knows patterns from what it was trained on, and that information can be out of date or wrong.

"AI is one single thing." No. "AI" covers many different programs. The one that blurs your video background has nothing to do with the one that writes poems.

"Only experts can use it." No. Many AI tools are built into apps you already have, and you talk to them in ordinary language.

Try this

1. Open the photo app on your phone and search for a word like "dog", "tree" or "food". The results are found by AI that recognises what is in each picture.
2. If you use an AI chatbot, ask it a question you already know the answer to, such as a fact about your home town. Notice what it gets right and anything it gets wrong.
3. Read our guides on using AI tools safely and AI for everyday tasks to make the most of it.