AI, machine learning & deep learning
Three terms used interchangeably that actually nest inside each other. Untangle them once and they stay clear.
People throw around "AI," "machine learning," and "deep learning" as if they're the same thing. They're not synonyms, they're nested circles: deep learning sits inside machine learning, which sits inside AI. Get this once and a lot of confusing conversations suddenly make sense.
The three circles
Artificial intelligence is the biggest circle, and the oldest. It's the broad goal: machines doing things that seem to require intelligence. That's it. A chess program from the 1980s following hand-written rules counts as AI. So does a modern LLM. The term is deliberately wide.
Machine learning is a circle inside AI. The key shift: instead of a person writing the rules, the system learns the rules from examples. You don't tell it "an email with these words is spam." You show it thousands of spam and not-spam emails, and it works out the pattern itself. (The next page makes this concrete.)
Deep learning is a smaller circle inside machine learning. It's machine learning done with many-layered neural networks, and it's what powers today's LLMs, image generators, and Agentforce. "Deep" just refers to those many layers. We'll meet neural networks in a couple of pages.
Why it's worth keeping straight
Here's the practical payoff. When someone says "AI" in 2026, they almost always mean deep learning specifically, the innermost circle, even though they're using the broadest word. Knowing that lets you ask the sharper question: which kind of AI are we actually talking about? A lead-scoring model and a chatbot are both "AI," but they live in very different parts of this picture, and they behave completely differently.
Rule of thumb: AI is the goal, machine learning is the method (learn from data), and deep learning is the specific method behind today's boom.
📝 Practice
Draw the three nested circles from memory: AI on the outside, machine learning inside it, deep learning in the middle. If you can sketch that on a whiteboard and say one sentence about each, you're ahead of most people talking about AI this week.
Predictive vs. generative AI
A bit of a trick question: both are prediction. Here's the real difference, with familiar Salesforce examples.
Data, the fuel
Models are built from data, and almost everything surprising about how they behave traces back to what they were fed. Where it came from, why diversity matters, and the strange problem of the internet starting to feed on itself.