Lesson 1

What is AI, Really?

Before we can understand ChatGPT, Claude, or any AI tool, we need to understand the building blocks. This lesson clears up the most common confusion: what's the difference between AI, machine learning, and deep learning?

After this lesson, you'll be able to:

  • ✓ Distinguish between AI, machine learning, and deep learning
  • ✓ Explain how machine learning differs from traditional programming
  • ✓ Identify the three types of machine learning

The Nesting Dolls of AI

People use "AI," "machine learning," and "deep learning" interchangeably. They're not the same thing. Think of them as nesting dolls:

Artificial Intelligence (AI)

The broadest concept. Any system that can perform tasks that normally require human intelligence. This includes everything from a chess program to Siri to ChatGPT.

Machine Learning (ML)

A subset of AI. Instead of being programmed with explicit rules, these systems learn patterns from data. You show it thousands of examples, and it figures out the rules on its own.

Deep Learning

A subset of ML that uses neural networks with many layers. This is what powers ChatGPT, image generators, and most modern AI breakthroughs.

Think of it this way: All deep learning is machine learning, and all machine learning is AI. But not all AI is machine learning. A rule-based spam filter that checks for specific keywords is AI, but it's not machine learning because it doesn't learn from data.

How Does Machine Learning Actually Work?

Traditional programming works like this: you give the computer rules and data, and it gives you answers.

Machine learning flips this. You give the computer data and answers, and it figures out the rules.

Traditional Programming

Rules + Data → Answers
"If email contains 'free money,' mark as spam"

Machine Learning

Data + Answers → Rules
"Here are 10,000 emails marked spam or not-spam. Find the rule."

This is why ML is so powerful. Instead of a human trying to write every possible rule, the machine discovers patterns that humans might never notice. But here's the catch: what it's actually finding are correlations in the training data. If the data is biased, so are the patterns.

The Three Types of Machine Learning

Supervised Learning
You give the model labeled examples (input + correct answer). It learns to predict answers for new inputs.
Example: Showing thousands of photos labeled "cat" or "dog" so it can classify new photos.
Unsupervised Learning
You give the model data without labels. It finds patterns and groups on its own.
Example: Grouping customers by purchasing behavior without telling it what the groups should be.
Reinforcement Learning
The model learns by trial and error, receiving rewards for good actions and penalties for bad ones.
Example: Training a robot to walk by rewarding it when it stays upright.
Analogy: Supervised learning is like learning with a teacher who tells you the right answers. Unsupervised learning is like exploring a new city without a map. Reinforcement learning is like learning to ride a bike — you fall, adjust, and eventually get it right.

Understanding these categories helps you evaluate AI tools critically. When someone says "our product uses AI," you can ask: What kind? Is it a rule-based system? A machine learning model trained on data? A deep learning model? Each has different strengths, limitations, and implications for how much you should trust its output.

Most of the AI making headlines today — ChatGPT, Claude, Midjourney, self-driving cars — is deep learning. That's what we'll focus on for the rest of this course.

Key Takeaway

AI is the broadest concept, machine learning is a specific approach within AI that learns from data, and deep learning is a powerful type of ML using multi-layered neural networks. The key shift: instead of programming rules, you show the machine examples and it figures out the rules.

Try This (Optional)

Pick three apps you used today — your email, your phone keyboard, a streaming service. For each, ask: is the smart part rule-based programming, or is it likely learning from data? "AI" covers a wider spectrum than the headlines suggest, and noticing the difference is the first useful skill.

Knowledge Check

A company feeds their AI system 50,000 customer emails that have been labeled as "complaint," "question," or "compliment." The system learns to automatically categorize new incoming emails. What type of machine learning is this?

Correct! The emails are labeled (complaint, question, compliment), so the model is learning from labeled examples. That's supervised learning.
Not quite. The key clue is that the emails are labeled — the model is given both the data and the correct answers. That's supervised learning.

Two startups both market "AI-powered" spam filters. Startup A wrote a list of keyword rules: if an email contains "free money," flag it. Startup B trained a model on millions of labeled emails to find patterns on its own. Which statement is most accurate?

Correct! AI is the broad umbrella — rule-based systems count. Machine learning is the narrower category that learns patterns from data, which is what Startup B is doing.
Not quite. Remember the nesting dolls: rule-based systems are AI (just not ML). Startup A's keyword filter is AI. Startup B goes further by learning from data — that's machine learning.
All Lessons Next: Neural Networks