The types of AI you'll actually meet

In this guide, you will learn how to tell the different types of AI apart, so the word stops feeling like one big blur.

Narrow AI and the AI of the films

Here is a distinction that clears up a lot of confusion. Almost every AI you have ever used is very good at one particular thing and useless at everything else. The spam filter that tidies your inbox cannot plan your holiday. The tool that recommends films cannot write you a poem.

This one-task-at-a-time kind is called narrow AI, and it is all we have today. It can be dazzlingly clever inside its lane and completely blank outside it.

The AI you see in films, the kind that thinks, feels, and does anything a person can, has a name too: general AI. It does not exist yet, and nobody knows for certain when or whether it will. So when a headline blurs the two, you can now tell which is real and which is imagination.

Learning from examples, not rules

Most modern AI has one thing in common: nobody sat down and wrote out every rule it follows. Instead, it was shown huge numbers of examples and worked out the patterns for itself, a bit like how a child learns what a dog is by seeing many dogs rather than by reading a definition. This approach of learning from examples is called machine learning.

That is why these tools can handle messy, real-world tasks that would be hopeless to spell out as strict rules. There is no rulebook for “what makes an email sound friendly,” but there are millions of friendly emails to learn from.

The tools you’re likely to meet

In day-to-day life, narrow AI turns up in a handful of familiar shapes.

The most common is the kind you type to and it types back, like ChatGPT or Claude. A tool you hold a conversation with is called a chatbot. It can draft, explain, summarise, and talk things through with you.

Then there are the image generators, such as Midjourney or DALL-E. You describe a picture in words and it paints one for you. Some tools now handle words, images, sound, and more all at once. AI that works across several kinds of content like this is called multimodal AI.

You will also meet quieter assistants stitched into apps you already use: the writing suggestions in your email, the smart replies on your phone, the voice assistant on your kitchen counter. Same family, different disguise.

Generative and predictive AI

One more split is worth knowing, because it explains what each tool is for.

Some AI creates brand-new things: fresh text, images, music, or code that did not exist before. This creative sort is called generative AI, and it is behind most of the tools making headlines today.

Other AI does not create; it predicts or sorts. It answers questions like “is this email spam?” or “how much will this house sell for?” You could call this the predictive sort. It is less flashy but quietly runs a great deal of the technology around you.

Try it yourself

The best way to feel the difference is to make one tool do two jobs. Open any AI chat tool and ask it to create something from scratch:

Write a short, cheerful four-line poem about a cat who has decided
the washing basket belongs to her now.

That is generative AI at work, inventing something new. Now give it a sorting task instead:

Here are three film reviews. Tell me which one is positive, which is
negative, and which is mixed, and explain how you can tell.
(Then paste or type three short reviews.)

Notice how the same tool shifts from creating to judging. Seeing both in one place makes the distinction stick.

A common first instinct

A common first instinct is to assume every AI can do everything, and then to feel let down when one falls flat. You will get more out of these tools if you match the task to the type: reach for a generative tool when you want something made, and remember that a tool built to sort or predict is not meant to chat. Picking the right shape for the job is half the skill.

Next steps

Now that you can tell the types apart, there is a practical choice waiting: should you use AI that lives on the internet, or AI that runs on your own device? That is exactly what we look at next in cloud versus local AI.