AI Glossary: Key Terms Explained Simply

Hallucination: when AI confidently states something false.
Grounding: giving AI source material so it answers from facts you provide.
Temperature: a setting that controls how random or predictable output is.
Bias: unfair patterns in AI output, often inherited from training data.

AI comes with a lot of jargon. Here are the terms you're most likely to see, explained simply.

Core concepts

Artificial intelligence (AI): computer systems that perform tasks we associate with human intelligence, such as understanding language or recognising images.
Machine learning: a way of building AI by training systems on examples rather than programming every rule.
Generative AI: AI that creates new content: text, images, audio, video or code.
Model: the trained AI system itself.
Large language model (LLM): a model trained on huge amounts of text to understand and generate language.

Prompt: the instructions or input you give an AI.
System prompt / custom instructions: background guidance that shapes every response.
Token: a chunk of text (often part of a word) that models read and write. Limits and pricing are often measured in tokens.
Context window: how much text the model can consider at once.
Multimodal: able to work with more than one type of input, like text and images.

Fine-tuning: further training a model on specific examples to specialise it.
RAG (retrieval-augmented generation): looking up relevant documents and giving them to the model before it answers.
Agent: an AI system that can plan and take multi-step actions using tools.
API: a way for software to connect to an AI model.

Diffusion model: the type of model behind many image generators, which turns noise into an image step by step.
Negative prompt: things you don't want in an image.
Seed: a number that controls the starting point of a generation, used to reproduce results.
Inpainting / outpainting: editing part of an image / extending it beyond its borders.