Lesson 1 of 9
What even is AI?
5 min read
'AI' gets stuck on everything from a chatbot to a spam filter to a sci-fi robot. So what actually is it — and where does ChatGPT fit?
One word, nested layers
Picture nested boxes. Artificial intelligence is the outer box: any machine doing things that seem to need human intelligence. Inside sits machine learning (it learns from examples, not hand-written rules), inside that deep learning (many-layered neural networks), and at the centre the LLMs you chat with. Every LLM is AI — but most AI you meet is narrow, good at just one thing.
AI is a nesting: LLMs are a small, specific slice near the centre.
Narrow now, not general
Today's systems are narrow: brilliant at one task, clueless outside it. The dream of one system that matches humans at everything — AGI — doesn't exist yet, and no one agrees when or if it will. When a headline says 'AI', it almost always means a narrow tool, not a mind.
We have many narrow tools, not one general mind.
'AI' is a marketing word as much as a technical one. When you hear it, ask: which narrow thing does this actually do?
The short version
- —AI ⊃ machine learning ⊃ deep learning ⊃ LLMs — nested, not synonyms.
- —Today's AI is narrow: great at one task, not general.
- —AGI (human-level, general) doesn't exist yet.
Which statement is true?
Continue in the app
Take the whole AI, Explained course — tracked
Get your personalized path, progress and streaks in the app — this lesson and every next one, in order.