avday How we use AI
We're a publication about AI that uses AI. Here is exactly how.
Where we use AI
- Research: finding and summarising sources, which we then read ourselves.
- First drafts: an AI model may produce a first draft from our editorial instructions.
- Editing help: tightening language, checking structure, suggesting headlines.
Where we don't
- Tests, benchmarks and screenshots are always our own. We never present AI-generated results as testing.
- Opinions are the named author's own.
- Quotes and interviews are never generated.
- Images that are AI-generated are labelled as such in the credit.
Human review
Nothing goes from an AI model straight to the site. Every draft arrives as a proposed change that a human editor must approve. For full articles, the editor checks every claim against a source, opens every link and adds something AI can't — code, a test, data, screenshots or an argued opinion. The reviewer is named on the article.
Briefs — short, fast takes on breaking news — get a quicker human check. They're labelled “Brief” and kept out of search engines for that reason.
Reviewers and sources
How we use AI is set out on this page. Every article names the editor who reviewed it and lists its sources. If you think we've got something wrong, please tell us.