Glossary

AI in plain language

The words you keep running into around AI, from prompt to deepfake, explained one by one in plain language. Click a word for the full explanation.

This is a glossary for people who aren't technical. Each word gets a short definition, an explanation with an example, and why it matters. The list keeps growing, every word checked by a human.

Search a word

AGI: AGI is the idea of an AI as broad and flexible as a human; it does not exist yet and is mostly a thing of the future.
AI: AI (artificial intelligence) is software that learns patterns from lots of examples, instead of following fixed rules.
AI agent: An AI agent is AI that does not just answer, but carries out steps on its own to finish a task, like looking something up or booking a meeting.
AI assistant: An AI assistant is an AI that helps with tasks in your work, such as writing, planning or summarizing, often built into a program you already use.
AI detector: An AI detector is a tool that tries to guess whether a text or image was made by AI; it is not reliable.
Algorithm: An algorithm is a fixed set of steps a computer follows to carry out a task or make a choice.
API: An API is a kind of socket that lets programs talk to each other; through an API, other software can use the AI.
Automating with AI: Automating with AI means having AI do recurring chores, so you keep time for the work that really matters.
AI phishing: AI phishing is scam messages made with AI, which makes them more convincing and more personal than before.

Bias: Bias is a skewed or one-sided picture an AI has picked up from its examples, which can make it respond unfairly or with prejudice.

Copyright and AI: Copyright and AI is about the question of who owns AI results, and whether an AI may use someone else's work.
Chatbot: A chatbot is a program you hold a conversation with in plain language: you type something, the bot answers.
ChatGPT: ChatGPT is a popular AI chatbot from the company OpenAI that lets you write texts, ask questions and summarize, all in plain language.
Cloud: The cloud means software and data run on a company's computers elsewhere, and you use them over the internet.
Context window: The context window is how much text an AI model can take in at once: your question plus everything you add to it.
Copilot: Copilot is Microsoft's AI assistant that works alongside you in programs like Word, Outlook, Excel and Teams.

Dataset: A dataset is a collection of data that belongs together, for example to train or test an AI.
Deepfake: A deepfake is a fake video, photo or voice made with AI that looks real but is not.

GDPR and AI: The GDPR is the privacy law that governs how you handle personal data; with AI that means: be careful what you put into a tool.
Gemini: Gemini is Google's AI, usable as a chatbot and woven into Google services like Gmail and Docs.
Generative AI: Generative AI is AI that makes new things: text, images, sound or code, instead of only recognizing things that already exist.

Hallucination: A hallucination is an answer an AI gives confidently, but that is made up or wrong.

LLM (language model): An LLM (large language model) is an AI model trained to understand and produce text.

Machine learning: Machine learning is how AI learns: by finding patterns in examples on its own, not from rules a person types in.
Model: A model is the trained program inside an AI tool that actually produces the text, the image or the answer.
Multimodal: Multimodal means an AI can handle more than one kind of input: not just text, but also images, sound or video.

Neural network: A neural network is the computing structure many AIs use to recognize patterns, loosely inspired by how brain cells work together.

Open-source model: An open-source model is an AI model that is freely available, so anyone may use it, inspect it and run it themselves.

Parameter: Parameters are the adjustable dials inside an AI model that get tuned during training; together they decide what the model 'knows'.
Privacy and AI: Privacy and AI is about the fact that what you type into an AI tool does not always stay private, and is sometimes used to train the model.
Prompt: A prompt is the instruction or question you give an AI tool.
Prompt engineering: Prompt engineering is the skill of phrasing your instruction to an AI so that you get a better result.

Summarizing with AI: Summarizing with AI is having AI cut a long text, email or meeting down to the key points and action items.

Token: A token is a small piece of text, often a word or part of a word, that a language model breaks text into.
Training: Training is the phase where an AI model learns: it sees an enormous number of examples and adjusts itself a little at a time.
Training data: Training data is all the examples an AI model was trained on: the texts, images or numbers it learned from.
Transcribing: Transcribing is turning spoken words into text; AI can type out a conversation or meeting automatically.
Translating with AI: Translating with AI is having text converted from one language to another, often more fluently and faster than before.

Rather have the whole Academy?

The glossary explains single words. The Academy takes you step by step, from the basics to confident use in your work. The first month is free.

Start free →Explore the Academy →

Last reviewed: 19 June 2026.