Definitions

These days, the term AI gets slapped onto just about anything that involves code. Marketing teams want us to believe that everything is powered by artificial intelligence. Your washing machine? AI. Your electric toothbrush? AI. Your toaster? You guessed it—AI.

But what is artificial intelligence, really?

To answer that, we need to strip away the word “artificial” and ask a more fundamental question: What does it mean to be intelligent?

What is Intelligence?

Consider this example:

Pedro, your friend who recently finished his master’s in physics, is standing at your apartment door, ready to celebrate his academic achievement. To mark the occasion, you thought it’d be fun to make some homemade pizza together. You’ve already picked up all the ingredients, so you let him in and head to the kitchen.

As you chat, you can’t help but admire Pedro’s brilliance. He’s the kind of person who can explain complex physics concepts with ease. Naturally, you joke that he should handle the cooking—after all, you got the ingredients, and you’re feeling a bit lazy. Pedro, always up for a challenge, agrees.

But as he fumbles through the spice rack, misreads the dough instructions, and debates whether mozzarella counts as “the white cheese,” it dawns on you: maybe Pedro isn’t quite the kitchen prodigy you expected. Sure, he’s brilliant at quantum mechanics—but can’t tell basil from oregano.

And that gets you thinking: What actually makes someone intelligent?

There’s no one-size-fits-all answer. Intelligence can be highly contextual. Are we talking about emotional intelligence? Social intelligence? Artistic talent? Raw analytical horsepower?

Is someone intelligent if they excel in one domain but struggle in others? Does it come down to learning speed, problem-solving ability, or adaptability? And how do we compare human intelligence to that of animals—or machines?

While there are many perspectives, one particularly influential definition comes from Marcus Hutter and Shane Legg:

“Intelligence measures an agent’s ability to achieve goals in a wide range of environments.”

Another, from psychologist Reuven Feuerstein, describes intelligence as:

“The unique propensity of human beings to modify their cognitive functioning to adapt to the changing demands of life.”

Both definitions emphasize adaptability and goal-oriented behavior—traits we generally associate with intelligence in both humans and machines.

So, put simply: Intelligence is the ability to adapt and achieve goals in a changing environment.

Now that we have a better idea of what it means to be intelligent, how does artificial intelligence fit into this picture?

What Is Artificial Intelligence?

Artificial Intelligence (AI) is a field within computer science and engineering dedicated to building machines that can behave intelligently—machines that attempt to replicate or simulate aspects of human cognitive function. That includes tasks like learning, planning, problem-solving, decision-making, and communication.

John McCarthy, one of the founding fathers of the field, famously described AI as:

“The science and engineering of making intelligent machines, especially intelligent computer programs.”

As you can probably guess, what counts as “intelligent” behavior depends on how we define intelligence in the first place—which is why the term AI can feel fuzzy or overused. AI encompasses a wide variety of subfields, including Machine Learning, Natural Language Processing and Computer Vision among others.

So, what exactly is machine learning? A machine that learns? Well, yes—but what does it actually mean to learn?

What is Learning?

Learning is a natural part of human life. It happens constantly and in many forms.

A child touches a hot stove and quickly learns not to do that again. An adult is betrayed by a friend and learns to be more cautious next time. At its core, learning involves a change in behavior. We do something a certain way, realize the result wasn’t what we expected, and adjust accordingly—so the next time something similar happens, we behave differently.

The number of times it takes to learn something varies. Touching a hot stove might only take one painful experience. Solving a complex math problem might take dozens of tries.

As with intelligence, the definition of learning also varies depending on the context. In their 2016 paper On the Definition of Learning, Hansbøl et al. describe learning as:

“A relatively permanent change in behavior brought about by practice or experience.”

In essence: learning is the process of adapting behavior through experience.

Now that we’ve covered what intelligence and learning are, we can finally bring everything together and ask:

What is Machine Learning?

Machine Learning (ML) is a subset of Artificial Intelligence (AI)

Machine Learning (ML) is a subset of AI focused on developing algorithms that can learn from data.

As Arthur Samuel famously put it:

“Machine learning is the field of study that gives computers the ability to learn without being explicitly programmed.”

In other words, machine learning allows machines to adapt their behavior based on new data—rather than relying entirely on hard-coded instructions.

Let’s go back to a simple example: imagine you work at Evilcorp and your job is to create a system that blocks inappropriate comments. One basic approach is to use a blacklist: check every comment against a list of offensive words and block anything that matches.

But what if people start getting creative—misspelling words, using symbols, or inventing new slang? Suddenly, your hardcoded filter needs constant manual updates.

With machine learning, however, the system can be trained on examples of inappropriate and appropriate comments. Over time, it learns patterns and improves its performance—without needing a developer to manually update the rules every time someone tries to get clever.

The Difference Between AI and ML

It’s common to hear the terms “AI” and “machine learning” used interchangeably—but they’re not the same.

Machine learning is a subset of artificial intelligence. Think of AI as the big picture, and ML as one powerful tool within it.

Not all AI involves machine learning. Some AI systems are based on manually defined rules and logic. These are sometimes called rule-based or hardcoded AI. They can behave intelligently, but only within very narrow constraints, and they don’t adapt to new data. If the problem changes, a human developer has to go in and change the code—just like in the Evilcorp example.

By contrast, machine learning systems learn from data and improve over time. They’re dynamic, flexible, and capable of adapting to new challenges.

Summary

• Intelligence is the ability to adapt and achieve goals in changing environments. This includes the ability to learn, plan, solve problems, make decisions, and communicate (through language for example).
• Artificial Intelligence aims to replicate this intelligence in machines.
• Machine Learning is one way to do that—by enabling machines to learn and adapt from data.

So next time someone tells you their fridge is “powered by AI,” you’ll know exactly what to ask: Does it actually learn anything? Or is it just another marketing gimmick?

Now that we know what machine learning is, let’s break down what it does and how it works.