Common AI terminology
As you learn about AI, you will encounter many terms. Here are some of the most common:
• Prompt — the instruction or question you give an AI • Token — a small piece of text the AI processes (roughly part of a word) • Model — the trained AI system itself (e.g. GPT-4) • Training — the process of teaching an AI using data • Inference — when the AI generates a response for you • Hallucination — when AI states something false confidently • RAG — Retrieval-Augmented Generation; connecting AI to your documents • Embedding — a way to represent meaning as numbers for search
Do not worry about memorising everything at once. Use the AI Dictionary to look up any term when you need a deeper explanation.
Terms in context
- •"Write a shorter prompt" = give clearer instructions
- •"This model is faster" = this version of the AI responds more quickly
- •"Check for hallucinations" = verify facts the AI stated
- •"We use RAG for our internal docs" = the AI searches company files before answering
Key points
- ✓Learn key terms gradually — you do not need to memorise everything
- ✓Prompt, token, model, and hallucination are especially common
- ✓The AI Dictionary explains each term in plain English
- ✓Lessons link to dictionary entries for deeper learning
Knowledge check
What is a "prompt" in AI?
Choose the best answer, then check your understanding.