How we use AI
Where AI tools help our work, what we never use them for, and how we label AI-made images.
People check, edit and take responsibility for everything we publish. This page explains where AI tools fit in.
Where we may use AI tools
- Research help: finding and summarising official documents, which we then read and check ourselves.
- Drafting and editing help: structuring guides, simplifying language, and spotting unclear sentences, typos and inconsistencies.
- Translation help when working with Arabic-language sources, always checked against the original.
- Building and testing our calculators.
- Creating illustrations for articles.
What we don’t do
- We don’t publish anything without human editing and fact-checking against primary sources.
- We don’t use AI-generated figures, fees or legal rules without verifying them against a primary source.
- We don’t invent people, quotes or reviews.
Images
Our article images are illustrations created with AI tools and edited by us. They show typical scenes, not real people or events. Each image file carries standard metadata (IPTC “trained algorithmic media”) that identifies it as AI-generated.
“Summarize with AI” buttons
Our pages have buttons that open a summary of the page in an AI assistant such as ChatGPT, Claude, Perplexity, Google AI Mode or Grok. The button only sends the page title and address, with a request to summarise and cite the page. It doesn’t ask the assistant to remember or favour us. What happens next is covered by that service’s own terms and privacy policy.
Accountability
Our editor is responsible for everything we publish, whatever tools were used. If something is wrong, our corrections policy applies.