AI words, explained simply.

The instruction you type into an AI tool. The clearer and more specific it is, the closer the result is to what you pictured.

The skill of writing and refining prompts so an AI model gives reliable, high-quality results.

AI that creates new content — text, images, audio, video or code — instead of only sorting or analysing existing data.

AI that understands and produces more than one kind of input — text, images, audio and video — in the same conversation.

An AI system that doesn't just answer — it plans steps and uses tools (browser, apps, code) to complete a task for you.

Creating a picture from a written description. The core feature of ChatGPT images, Gemini, Midjourney and Flux.

Giving the AI an existing photo plus instructions, so it transforms that image instead of starting from nothing.

Generating a short video clip from a written description — the technology behind Veo, Sora and Kling.

The main technology behind AI images and video: the model starts from random noise and gradually refines it into a picture that matches your prompt.

Telling an image model what you don't want — extra fingers, blur, text, watermarks — so it avoids those things.

The number that sets the random starting point of an AI image. Same prompt + same seed gives (nearly) the same image.

The width-to-height shape of an image or video, like 9:16 for reels, 4:5 for Instagram posts or 16:9 for YouTube.

Using AI to enlarge an image and add believable detail, so a small picture looks sharp when printed or viewed full-screen.

Editing part of an image with AI (inpainting) or extending it beyond its edges (outpainting).

A small add-on file that teaches an image model one specific face, product or art style without retraining the whole model.

The type of AI behind chatbots like ChatGPT, Gemini and Claude, trained on huge amounts of text to predict and generate language.

The small chunks of text an AI model reads and writes — roughly a word or part of a word. Limits and pricing are counted in tokens.

How much text an AI model can keep in mind at once — your conversation, uploaded files and its own replies combined.

An AI model that works through a problem step by step before answering, trading speed for accuracy on hard questions.

Getting an AI to show its working step by step, which often makes its final answer more accurate.

A setting that controls how predictable or creative an AI's output is — low for consistent answers, high for variety.

A technique where the AI first looks up relevant documents and then answers using them, so replies are grounded in real sources.

Further training an existing AI model on a smaller, specific dataset so it gets better at one style or task.

An AI model whose weights are published so anyone can download, run and modify it — often for free.

A way for apps and developers to use an AI model from their own code, usually paid per use (per token or per image).

When an AI confidently states something false — a fake fact, quote, source or statistic.

AI-generated or AI-altered video, audio or images that make a real person appear to say or do something they didn't.

An invisible or visible marker that shows an image, video or text was made by AI, so it can be identified later.