The AI Glossary — Plain Words for Every Term

The AI Glossary — Plain Words for Every Term You Will Meet

AI Universe City · The AI Agency Learning Community · Free Classroom · 2026-09-16

What this is: a reference. Eight short entries explaining the words that get used constantly in AI, each in plain language with an analogy.

How to use it: do not read this front to back. Dip into it when a term comes up and you are not sure what it means. Every entry takes three minutes.


Why this exists

AI is full of words that sound harder than the ideas behind them. Token, inference, context window, quantization, prompt injection. Read a product page or a forum thread and you will meet all of them in one paragraph.

Once you know the words, the conversation stops being intimidating. That is the whole purpose of this glossary. Nothing here is deep, and none of it requires a technical background.

The entries

1. What a Model Actually Is — the basic unit, and what “training” really means.

2. Talking to Models: Prompts and Context — the two words you will use most.

3. Running AI: Tokens, TPS, VRAM and RAM — how to read a spec sheet without glazing over.

4. When AI Acts: Tools, Agents and MCP — where AI stops talking and starts doing.

5. Making It Cheaper: Quantization, MoE and Distillation — the tricks that let big models run on small machines.

6. Voice AI Vocabulary — latency, TTS, STT, and why voice is harder than text.

7. Automation Vocabulary — triggers, webhooks, and wiring systems together.

8. The Limits: Hallucinations, Bias and Privacy — the three failure modes, and what never to paste into a chat box.

Three habits worth keeping

These come from the final entry, and they are the most useful thing in the glossary.

  1. Facts get a source check. Every time.
  2. Anything that judges a person gets a human check. Every time.
  3. Private data stays out of the chat box. Every time.

Where this fits

The glossary stands on its own. If you prefer to learn by doing, go straight to AI Foundations and come back here when a word stops you.

If you are preparing to work through the paid Academy track, reading entries 3, 5 and 7 first will make the material land much more easily.


Found a term that is not here? Post it in Ask & Answer and it will be added to this glossary.