Skip to content
AI Digested
GuideBeginner3 min read

AI vs machine learning vs deep learning: what's the difference?

AI, machine learning, deep learning, generative AI: people use these words as if they mean the same thing. Here's how they fit together, in one simple picture.

First published · AI Digested

You'll hear "AI", "machine learning", "deep learning" and "generative AI" used almost interchangeably, sometimes in the same sentence. They are related, but they aren't the same. The easiest way to picture them is as circles inside circles.

The big circle: artificial intelligence

Artificial intelligence is the broadest term. It covers any technique that makes a computer do something we'd normally say needs intelligence: understanding language, recognising images, making decisions, playing games.

AI includes some very old ideas that don't learn at all. A 1980s "expert system" that diagnoses car faults by following thousands of if-then rules written by mechanics is AI. So is a video game character that chases you using a fixed set of rules.

Inside it: machine learning

Machine learning is one way of building AI, and today it's by far the most important one. Instead of writing the rules, you give the computer examples and let it learn the rules itself.

A spam filter is the classic example. Nobody writes a complete list of spam words. The filter is shown millions of emails that people marked as spam or not spam, and it learns which patterns predict spam, including ones no human would have noticed.

Machine learning covers many techniques. Some are quite simple statistics that have been used for decades, such as predicting a house price from its size and location.

Inside that: deep learning

Deep learning is one particular kind of machine learning. It uses neural networks, layers of simple connected calculations loosely inspired by brain cells. "Deep" just means the network has many layers.

Deep learning needs a lot of data and a lot of computing power, which is why it only took off around 2012, when large datasets and powerful graphics chips became available. Since then, it has powered almost every headline-making AI: face recognition, speech recognition, self-driving car vision, AlphaGo and every modern chatbot.

Overlapping them all: generative AI

Generative AI describes what a system does rather than how it's built: it generates new content, such as text, images, audio, video or code.

Today, generative AI is almost always built with deep learning. Large language models, which power chatbots like ChatGPT, Claude and Gemini, are a type of generative AI that works with text.

One example, all the terms

Take a chatbot like ChatGPT:

Question Answer
Is it AI? Yes: it does something that seems to need intelligence.
Is it machine learning? Yes: it learned from examples rather than hand-written rules.
Is it deep learning? Yes: it is a very large neural network with many layers.
Is it generative AI? Yes: it writes new text.
Is it a large language model? Yes: specifically, it's an app built on one.

Now take your bank's fraud alerts. They are AI and machine learning, sometimes deep learning, but not generative AI: they flag suspicious payments rather than create anything.

Why the words get muddled

Partly marketing. "AI" sells, so almost any software with a clever feature now gets the label. Partly because the field moves fast and the latest breakthrough tends to borrow the biggest word available.

When you see "AI-powered" on a product, a fair question is: what does it actually learn, and from what data? A good product can answer that.

Keep going

Common questions

Is all AI machine learning?

No. Some AI follows rules written by people, such as older chess programs or simple customer service bots with fixed menus. But nearly all of the AI you hear about today uses machine learning.

Is ChatGPT machine learning or deep learning?

Both. ChatGPT is built on a large language model, which is a deep neural network. Deep learning is a type of machine learning, so it is machine learning too.

What is data science?

Data science is the broader job of getting useful answers from data. It uses statistics, charts and often machine learning. It overlaps with AI but isn't the same thing.

Spotted a mistake or something out of date? Tell us and we will fix it. Read how we research and write our articles.