So, here's the thing. A few months ago, I kept hearing these four words everywhere — AI, machine learning, deep learning, generative AI — and honestly? I just nodded along like I knew what they meant. I didn't. Not really.
Then I actually sat down and learned them properly, and it turns out they're way simpler than people make them sound. So here's my attempt at explaining them the way I wish someone had explained them to me. No scary maths, no fifty-page research papers. Just the actual ideas.
These four terms are NOT the same thing. They're more like Russian dolls — each one fits inside the bigger one:
Once I saw it like that, everything else clicked into place.
Artificial intelligence is basically any system that does something we'd normally say requires human thinking. That's it. That's the definition.
And here's the part that surprised me — AI is old. Like, really old. The chess computer that beat Kasparov in 1997? That's AI. The enemy characters in old video games that chase you around? Technically AI. Most of that stuff doesn't "learn" anything. A programmer just wrote a bunch of rules, like "if the player is close, move towards them."
So AI ≠ robots taking over the world. Most AI is honestly pretty boring. It's spam filters, autocorrect, and the thing that recommends which video plays next.
This is where computers stopped following rules and started finding them.
The classic example: a spam filter. The old way, a programmer would write rules like "if the email contains the word FREE!!! in all caps, it's spam." But spammers just change their spelling and the rules break. It's a never-ending game of cat and mouse.
Machine learning flips it around. Instead of writing rules, you show the computer thousands of emails — some labelled "spam," some labelled "not spam" — and it figures out the patterns by itself. Nobody tells it what to look for. It just notices that certain words, senders and formats tend to show up in spammy emails.
I like to think of it like this: normal programming is giving someone a recipe. Machine learning is giving someone a thousand cooked meals and letting them figure out the recipe.
The thing it builds is called a model, and the process of building it is called training. You'll hear those two words constantly.
Deep learning is a specific kind of machine learning that uses something called a neural network — a structure loosely inspired by how brains work. (Very loosely. Anyone who says "it works just like the human brain" is overselling it.)
Picture it as layers of tiny decision-makers stacked on top of each other. Each layer looks at the data, picks out some pattern, and passes it on to the next layer. Early layers notice simple stuff; later layers combine that into complicated stuff.
The example that made it click for me: recognising a face in a photo.
Nobody programmed "this is what an eye looks like." The network worked it out from looking at millions of photos. It's called "deep" learning because there are many layers — that's literally where the name comes from. I was expecting something more profound, but nope, it's just... lots of layers.
The catch? Deep learning is hungry. It needs enormous amounts of data and serious computing power. That's partly why it only really took off in the last decade or so — we finally had the data (thanks, internet) and the hardware (thanks, gaming graphics cards) to feed it.
Okay, this is the one that's everywhere right now. Chatbots, image generators, those tools that write essays or make songs — that's generative AI.
Most machine learning is about sorting things: is this email spam or not? Is this a cat or a dog? Generative AI is about making things: writing a paragraph, drawing a picture, composing music.
Here's the slightly mind-bending part: a chatbot like ChatGPT is, at its core, doing something almost disappointingly simple. It's predicting the next word. Over and over. "The cat sat on the..." — what word probably comes next? "Mat." Then it predicts the word after that. And the next one.
But here's the thing — when you train that "next word predictor" on basically a huge chunk of the internet, the predictions get so good that it can hold a conversation, write poetry, explain quantum physics and help you debug code. All from "guess the next word." I still find that kind of wild.
Image generators work on a similar idea. They're trained on millions of images and learn what things tend to look like, so when you type "a golden retriever wearing sunglasses on a skateboard," it can generate something it has never seen before, because it understands the patterns of dogs, sunglasses and skateboards.
Fair question. Here's my honest take.
We're going to grow up in a world where these tools are everywhere — in our jobs, our apps, probably our homework (no comment). You don't need to become an AI researcher. But understanding the basics — even just what's in this post — means you can actually tell the difference between what's real and what's hype.
And there's a lot of hype. Generative AI is genuinely impressive, but it also makes stuff up with total confidence, reflects the biases in its training data, and doesn't "understand" anything the way we do. Knowing that is just as important as knowing what it can do.
Four dolls, one inside the other. That's genuinely most of what you need to know to follow any conversation about this stuff.
If you made it this far, thanks for reading! I'm planning to write more of these as I learn — next one's probably going to be about how training actually works, because that's the part that broke my brain the most (in a good way). Written by a student, for students. If I got anything wrong, that's on me — I'm learning too.