⚠️ Affiliate Disclosure: This post is not an affiliate associated review. It may become one in the future and this disclosure will reflect that. This does not affect my reviews and I only recommend tools I genuinely believe in.

📅 Last updated: 31 October 2026

What Is This AI Glossary?

This AI glossary explains the terms you will meet when using or reading about AI tools, in plain English and without maths. Each definition stands on its own, and where I have a full guide on a term, I link to it.

It is written for people using AI at work or at home, not for engineers — so it covers the words that actually turn up in product pages, pricing tables and news stories.

AI Terms: A

Agentic AI

The approach of building AI that acts with some independence: planning steps, using tools and making decisions towards a goal. See AI agent.

AGI (artificial general intelligence)

A hypothetical AI that could match humans across almost any intellectual task, rather than excelling at specific ones. No AI today meets that bar, and experts disagree about when, or whether, one will.

AI (artificial intelligence)

Computer systems that perform tasks normally associated with human intelligence, such as understanding language, recognising images or making decisions. Most AI you use today is machine learning.

AI agent

AI that takes actions to complete a goal — using tools like email, a browser or a calendar — rather than just answering questions. My guide: What Is an AI Agent?

AI credits

A unit many AI tools use to meter usage. Each action, image or block of text costs credits, which usually stand in for the tool’s underlying computing or token costs.

AI detector

Software that estimates whether text was written by AI. Detectors are unreliable and should never be treated as proof. My guide: How to Tell If Something Was Written by AI

AI hallucination

A confident AI answer that is false or invented, such as a fake source or statistic. My guide: AI Hallucinations

API

Application programming interface: a way for software to talk to other software. AI companies sell access to their models through APIs, usually billed per token, and many AI tools are built on them.

AI Terms: B – C

Bias (in AI)

Systematic unfairness in AI output, usually inherited from the human-written data a model learned from.

Chatbot

An AI you talk to in a conversation, such as ChatGPT, Claude or Gemini. Modern chatbots are built on large language models.

Context window

The maximum amount of text, measured in tokens, an AI model can consider at once — including your messages, uploaded files and its own replies. When a conversation outgrows it, earlier material drops out of view.

AI Terms: D – F

Deep learning

A type of machine learning that uses large neural networks with many layers. It is the technique behind modern language, image and voice AI.

Deepfake

Realistic AI-generated or AI-altered video, audio or images of a real person, made to look genuine.

Diffusion model

The kind of AI behind most image and video generators. It learns to turn random visual noise into a picture step by step, guided by your prompt.

Fine-tuning

Further training an existing AI model on a smaller, specialised set of examples so it performs better at a particular task or style.

Foundation model

A large, general-purpose AI model trained on broad data that other products are built on. GPT, Claude, Gemini and Llama are foundation models.

AI Terms: G – L

Generative AI

AI that creates new content — text, images, audio, video or code — rather than only analysing existing data.

GPT

Generative pre-trained transformer: the name of OpenAI’s family of large language models, which power ChatGPT.

Guardrails

Rules and filters that stop an AI producing harmful, unsafe or off-topic output.

Inference

The moment an AI model is used to produce an answer, as opposed to when it is trained. Most AI running costs are inference costs.

Knowledge cut-off

The date after which a model has no training data. Without web search, it knows nothing that happened later.

LLM (large language model)

AI trained on vast amounts of text to understand and generate language; the technology behind ChatGPT, Claude and Gemini. My guide: What Is an LLM?

AI Terms: M – O

Machine learning

The branch of AI where systems learn patterns from data rather than following hand-written rules.

MCP (Model Context Protocol)

An open standard that lets AI assistants connect to other apps and data sources, such as your files, calendar or business tools, in a consistent way.

Model

The trained AI system itself. An AI tool or app is the product you use; the model is the engine inside it.

Multimodal AI

AI that can work with more than one type of input or output — for example, reading an image and answering questions about it in text.

Neural network

A computing system loosely inspired by the brain, made of layers of connected units that adjust as they learn. The building block of deep learning.

NLP (natural language processing)

The field of AI concerned with understanding and generating human language. LLMs are the current state of the art in NLP.

Open-weights model

An AI model whose trained parameters are published so anyone can download, run and adapt it, such as Meta’s Llama. Often loosely called open-source AI.

AI Terms: P – R

Parameters

The internal settings a model adjusts during training. Model size is often quoted in parameters, typically billions.

Prompt

The instruction or question you give an AI. My guide: How to Write AI Prompts That Actually Work

Prompt engineering

The practice of writing and refining prompts to get better results. For most people, it simply means giving clear instructions and good context.

RAG (retrieval-augmented generation)

A technique where an AI looks up relevant documents or web pages first and answers from them, which reduces made-up answers and allows citations.

Reasoning model

An AI model designed to work through a problem in steps before answering, which improves results on maths, logic and planning at the cost of speed.

AI Terms: S – Z

Synthetic content

Text, images, audio or video generated by AI. New EU rules require much of it to be marked as AI-generated; see my guide to AI watermarking.

System prompt

Hidden instructions given to an AI before your conversation starts, setting its role, tone and rules. Custom instructions in chat apps work in a similar way.

Temperature

A setting that controls how predictable or varied an AI’s output is. Low temperature gives consistent answers; high temperature gives more creative, less reliable ones.

Token

The small chunk of text, roughly three-quarters of a word, that AI models read and write. Tokens determine usage limits and pricing. My guide: What Is a Token in AI?

Training data

The text, images or other material an AI model learns from. Its quality and coverage shape everything the model can and cannot do.

Transformer

The AI architecture, introduced by Google researchers in 2017, that almost all modern language models are built on.

Frequently Asked Questions

What are the most important AI terms to know?

For everyday use: AI, generative AI, large language model (LLM), prompt, token, context window, AI hallucination and AI agent. Those eight explain most of what you will read about AI tools and their pricing.

What is the difference between AI, machine learning and deep learning?

AI is the broad goal of machines doing intelligent tasks. Machine learning is the main way that is achieved today, by learning from data. Deep learning is a type of machine learning using large neural networks, and it powers modern chatbots and image generators.

What is the difference between narrow AI and AGI?

Narrow AI is built for specific tasks, like writing text or recognising faces, and is all that exists today. AGI, artificial general intelligence, would match humans across almost any intellectual task. It remains hypothetical.

What is generative AI in simple terms?

Generative AI is AI that makes new things — writing, pictures, music, video or code — from a prompt you give it. ChatGPT, Midjourney and Suno are all generative AI tools.

What type of AI is ChatGPT? Is it AI or machine learning?

Both. ChatGPT is generative AI built on a large language model, OpenAI’s GPT. That model was made with machine learning, specifically deep learning: it learned patterns of language from huge amounts of text rather than following hand-written rules. So ChatGPT is an AI tool, machine learning is how it was built, and “large language model” is the precise name for what it is. My guide: What Is an LLM?

What are the types of AI?

Lists of “the 7 types of AI” combine two ways of grouping it. By capability: narrow AI, general AI (AGI) and superintelligence. By how it works: reactive machines, limited-memory AI, theory-of-mind AI and self-aware AI. Only narrow, limited-memory AI exists today, and that includes every chatbot and AI tool you can use. The rest are theoretical.

James Huntington
James Huntington
AI Tools Expert

James is an SEO and content specialist with over 12 years’ experience, and tests AI tools so everyday people don’t have to. WhatAITools.co.uk covers honest, no-jargon reviews to help you find the right tool for your needs.

LinkedIn · Follow on 𝕏 (Twitter / X)


Tried this tool?

Share your experience below — your review helps other everyday people make better decisions.

Leave a Reply

Your email address will not be published. Required fields are marked *