AI Glossary (Plain English)

Core terms

Artificial intelligence (AI) - Software that performs tasks that usually need human judgment, often by finding patterns in data.

Machine learning - AI that improves from examples instead of hand-written rules.

Large language model (LLM) - A model trained on huge amounts of text to predict the next word (token). Powers ChatGPT, Claude, and similar tools.

Token - A small chunk of text the model reads and generates. Roughly ¾ of a word in English, but varies.

Prompt - The instructions and context you give a model. Quality in → quality out.

Hallucination - A fluent but false model answer. Always verify important facts.

Fine-tuning - Extra training on your examples to steer behaviour. Expensive; often unnecessary if prompting + RAG work.

RAG (retrieval-augmented generation) - Fetching your documents at question time and feeding them to the model so answers stay grounded.

Embedding - A numeric fingerprint of meaning used to search similar text.

Agent - An LLM in a loop that can call tools (search, code, APIs) to complete multi-step tasks.

Transformer - The architecture behind most modern LLMs; uses attention to weigh which context matters.

Go deeper

• Content library
• Blog
• Free session - experience these ideas live