Module 1: What Is AI, ML & Deep Learning?
Lesson 1.1 of 40

AI vs. ML vs. Deep Learning
vs. Data Science

18 min 📚 Beginner 🧠 Concepts Only
Section 1

Why These Terms Get Confused

Open any tech news article and you'll see “AI,” “machine learning,” “deep learning,” and “data science” used as if they're interchangeable. They're not. They're nested inside each other, like Russian dolls — and knowing exactly where one ends and the next begins will save you from a lot of confusing job postings, course descriptions, and conversations.

Here's the short version before we slow down: Artificial Intelligence is the broad goal — making machines do things that seem intelligent. Machine Learning is one approach to that goal — teaching machines by showing them examples instead of writing explicit rules. Deep Learning is one technique within machine learning — using layered neural networks specifically. Data Science is a different, overlapping discipline focused on extracting insight from data, which often uses ML as a tool but also includes plenty of work that has nothing to do with AI at all.

💡
If you remember nothing else from this lesson: ML is a subset of AI. Deep Learning is a subset of ML. Data Science overlaps with both but is its own field. Everything else in this lesson is detail on top of that one sentence.
Section 2

The Nesting, Visually

Picture four concentric circles. The outermost is AI — the biggest, vaguest category. Inside it sits ML. Inside ML sits Deep Learning. Data Science sits partially overlapping with AI and ML, but also extends outside both circles into territory that's just “working with data,” no learning algorithms involved.

🧠
Artificial Intelligence
The umbrella goal: any technique that makes a machine behave in a way that seems intelligent. Includes ML, but also rule-based systems, search algorithms, and symbolic logic that involve no "learning" at all.
📊
Machine Learning
A specific approach to AI: instead of programming explicit rules, you show the system examples and it learns patterns from them. This entire course track lives mostly here.
🔗
Deep Learning
A specific technique within ML: using neural networks with many layers ("deep" = many layers) to learn very complex patterns, especially from images, audio, and text.
📊
Data Science
A broader practice of extracting insight from data — statistics, visualization, business reporting, and yes, often ML too. Not every data scientist builds ML models; not every ML engineer does data science.
Section 3

Telling Them Apart With Real Examples

Abstract definitions only get you so far. Here's how the same four categories show up in things you've actually used:

ExampleCategoryWhy
A thermostat that follows "if temp < 68°F, turn on heat" AI (rule-based) It behaves "intelligently," but it's a hand-written rule, not learned from data. No ML involved.
A spam filter trained on thousands of labeled emails Machine Learning It learned the pattern "what spam looks like" from examples, rather than being told explicit rules.
Face recognition unlocking your phone Deep Learning Specifically uses a neural network with many layers to learn facial features directly from pixel data.
A quarterly sales report with charts and trend analysis Data Science (no ML) Pure statistics and visualization. No model is learning anything — it's analysis, not prediction.
Quick test for yourself: ask "did this system learn its behavior from examples, or was the behavior written explicitly by a person?" If it learned from examples, you're in ML territory. If a human wrote the exact logic, it's AI without ML.
Section 4

Where This Course Track Fits

This 4-tier track is built to walk you through all of this in order, not all at once:

Tier 1
Foundations
Math + first ML models
Tier 2
Classical ML
Trees, SVMs, clustering
Tier 3
Deep Learning
Neural networks, transformers
Tier 4
Modern AI
LLMs, GenAI, MLOps

Notice that "Deep Learning" doesn't show up until Tier 3 — on purpose. Most real-world ML (fraud detection, recommendation systems, forecasting) is built with the classical techniques in Tier 2, not deep learning. Skipping straight to neural networks, which is what a lot of online content pushes you toward, means missing the tools that actually run most production systems today.

⚠️
Common misconception to drop now: "Deep Learning" is not a more advanced version of "Machine Learning" that replaces it. It's one specific technique, best suited to certain problems (images, audio, text) and often worse than simpler methods on smaller, structured datasets — which is most business data.
✅ Quick Check — Lesson 1.1
1. Which statement correctly describes the relationship between these fields?
2. A thermostat that turns on heat below a hand-written temperature threshold is an example of:
3. Why does this course cover classical Machine Learning (Tier 2) before Deep Learning (Tier 3)?
🎉 Lesson 1.1 complete! Next: what AI is actually used for today.
🗒 Cheat Sheet 📝 Worksheet