"AI" is one of those phrases that's been stretched so far it barely means anything. Your toaster is "AI" now. So let's be specific.
The short version
AI, short for artificial intelligence, is any computer system that does something we used to think needed a human brain. Recognising a face. Translating a language. Writing an essay. Suggesting what to watch next.
That's existed for decades. The reason AI feels different in 2026 is because of one specific breakthrough: Large Language Models, or LLMs.
Why LLMs changed the game
Older AI was narrow. A spam filter was great at spotting spam and utterly useless at anything else. Google Translate could translate, but it couldn't summarise.
LLMs are general. The same model that writes a birthday poem can also:
- Draft a legal email.
- Explain quantum physics to a ten-year-old.
- Debug a spreadsheet formula.
- Roleplay as a job interviewer so you can practise.
One tool. Thousands of tasks. That's what's new.
The bit that catches people out
Here's the thing no one tells you at first. LLMs don't "know" things the way a search engine does. They've read a huge amount of text and learnt the patterns of language. When you ask a question, they generate an answer one word at a time based on what's statistically most likely to come next.
That's a superpower, and it's also the reason they sometimes make things up. The industry calls this "hallucinating". We'll dig into it properly in Lesson 5: Why AI gets things wrong. For now, just remember:
An LLM is a very, very good guesser. It is not a database.
What counts as "AI" in this course
When we say "AI" in this course, we almost always mean LLM-powered chat tools. ChatGPT, Gemini, Claude, and a few friends. Those are the tools that matter to you on a Tuesday afternoon when you've got a newsletter to write.
Next lesson, we'll meet the main ones and figure out which is best for what.