Generative AI is a type of artificial intelligence that creates new content: an email, an illustration, a song, a video clip or a piece of code. You describe what you want in plain language (a prompt), and the AI produces it in seconds.
How does it work?
Generative models are trained on enormous amounts of existing data, such as text, images or audio. During training they learn patterns: how words follow each other, what a "sunset over the ocean" tends to look like, how a melody resolves. When you give a prompt, the model uses those patterns to generate something new that fits your request.
Text tools like ChatGPT, Claude and Gemini are built on large language models, which predict text piece by piece. Many image and video tools use diffusion models, which start from random noise and gradually shape it into a picture that matches the prompt.
What it's good at
- First drafts of writing, from emails and articles to scripts
- Summarising and explaining complex information
- Brainstorming ideas, names and angles
- Creating images, music and video concepts quickly
- Helping with code, spreadsheets and data
Where it falls short
- It can state false information confidently ("hallucinations").
- It may reflect biases in its training data.
- It doesn't truly understand your situation unless you explain it.
- It can't replace professional judgement for legal, medical or financial decisions.
How to get started
Pick one everyday task, such as writing an email, summarising a document or planning a week, and try it with an AI assistant. Be specific about what you want, review the result and ask for changes. Browse the prompts on this site for ready-made starting points.