Three Terms, One Family

If you've spent any time reading tech news, you've almost certainly seen "AI," "machine learning," and "deep learning" used as if they mean the same thing — or, equally confusing, as if they're three completely separate inventions competing with each other. Neither is true. The real relationship is simpler: they nest inside one another like Russian dolls.

Analogy Think of it this way: AI is the whole country. Machine learning is one large state inside that country. Deep learning is a specific city inside that state. Every city is in the state, every state is in the country — but the country contains plenty of other states too.

Once you see the nesting-doll picture, the three terms stop being confusing and start being genuinely useful. Let's unpack each layer in plain English.

Layer 1: Artificial Intelligence — The Broadest Goal

Artificial intelligence simply means getting a machine to do something that would normally require human-like thinking. That definition is deliberately wide. It covers a chess program from the 1980s that follows hand-coded rules just as much as it covers a modern chatbot that writes poetry.

Early AI relied entirely on rules written by humans: "if the user says X, reply with Y." These systems were clever, but they broke the moment a situation wasn't already covered by a rule. You can explore this history in Lesson 1: What Is AI? — it's a great place to start if any of this feels brand new.

Layer 2: Machine Learning — Learning From Examples

Machine learning (ML) is the approach where instead of a programmer writing every rule by hand, the system figures out the rules itself by studying lots of examples. Show it thousands of cat photos and thousands of non-cat photos, and it learns what makes a cat a cat — without you ever typing "look for pointy ears."

Analogy Teaching a child to recognise dogs by showing them hundreds of dogs is machine learning. Writing a rulebook that says "four legs + fur + barks = dog" is old-school AI. The child's approach handles a chihuahua and a Great Dane equally well; the rulebook struggles the moment it meets a wolf.

ML works by finding statistical patterns in data. It's why your email app gets better at catching spam over time, why Netflix knows you'll probably enjoy a certain documentary, and why your phone can unlock with your face. To go deeper, Lesson 2: How Machines Learn walks through how this learning process actually works — no maths degree required.

ML comes in different flavours too. In supervised vs unsupervised learning, you'll learn the key difference between training a model with labelled examples versus letting it find its own structure in the data.

Layer 3: Deep Learning — ML With Many-Layered Networks

Deep learning is a particular technique within machine learning. It uses structures called neural networks that are loosely inspired by the way neurons connect in a brain. The "deep" in deep learning refers to the many layers stacked on top of each other inside these networks — a shallow network might have two or three layers; a deep one might have hundreds.

Each layer learns to recognise increasingly abstract features. In an image recognition network, the first layer might spot edges, the next layer spots shapes built from those edges, and later layers spot eyes, noses, or whole faces. The deeper you go, the more sophisticated the concept the network has learned to see.

Analogy Imagine a team of specialists reviewing a photo. The first specialist says "I see horizontal and vertical lines." They pass notes to the next specialist who says "those lines form a circle and two triangles." That specialist passes notes onward, and eventually someone says "that's a face." Each layer builds on the layer before it — that's depth.

Deep learning only became practical in the 2010s when two things arrived at the same time: massive datasets (think billions of labelled images on the internet) and powerful GPU hardware cheap enough to train huge networks in reasonable time. Without both, deep learning is just a good idea that doesn't work fast enough to be useful.

Deep learning is behind the most dramatic AI breakthroughs of recent years — image generation, real-time translation, speech recognition, large language models. Lesson 6: Deep Learning & Vision lets you actually play with this, and our article on What Is a Neural Network? explains the building blocks from scratch.

Side-by-Side Comparison

Term What it means Everyday example
Artificial Intelligence Any technique that lets a machine perform tasks requiring human-like intelligence — including rule-based systems and ML A GPS app that finds the fastest route to avoid traffic
Machine Learning A subset of AI where the machine learns patterns from data instead of following hand-coded rules A spam filter that improves automatically as it sees more emails
Deep Learning A subset of ML using multi-layered neural networks; especially powerful for images, audio, and text Your phone recognising your face in different lighting conditions

Which Term Should I Use?

In practice, people use "AI" as the catch-all term in everyday conversation, and that's fine. When you want to be more precise — for example, when describing how a system was built — reach for the more specific term. If the system learned from data, say "machine learning." If it used a neural network with many layers, say "deep learning." If it followed hand-coded rules, just call it "AI" or "a rule-based system."

Key takeaway All deep learning is machine learning. All machine learning is AI. But not all AI is machine learning, and not all machine learning is deep learning. When in doubt, zoom out: AI is always the safest, most inclusive term.

Frequently Asked Questions

Is deep learning the same as AI?

Not quite — deep learning is a specific technique that lives inside the broader AI umbrella. Think of AI as the goal (intelligent machines) and deep learning as one powerful method for reaching that goal. Plenty of AI systems — like a route planner or a game-playing program from the 1990s — use no deep learning at all.

Is all AI machine learning?

No. Machine learning is one of the most popular approaches to AI right now, but AI also includes rule-based expert systems, search algorithms, logic engines, and more. A classic chess engine that calculates moves using hand-coded rules is AI — but it doesn't learn from data, so it isn't machine learning.

Do I need to know maths to understand this stuff?

Not to get started. The concepts of AI, ML, and deep learning are intuitive and can be understood through analogies and interactive examples — which is exactly what this course is built around. If you later want to build your own models from scratch, some maths (statistics, linear algebra) becomes helpful, but you can go a very long way just by understanding the ideas and experimenting with tools.

Key takeaway Ready to explore further? The best next step is to start with the foundations and build up from there. Start with Lesson 1 →