Transparency
AI & content policy
We use AI as a newsroom tool, under human control, against attributed sources — never as an unsupervised writer.
Where AI is used
AI models assist with summarising and paraphrasing attributed source material, drafting headline and standfirst options, suggesting metadata and structured data, classifying stories into our taxonomy, and generating illustrative imagery when no licensed photograph is available.
Where AI is never used
AI does not decide whether a claim is true, does not invent quotations, statistics, named individuals or company statements, and does not publish anything on its own. Any output that cannot be traced to the underlying source is rejected before publication.
Labelling
Articles produced with AI assistance state that fact and show their verification status and full source attribution at the foot of the story. Illustrative images generated by AI are credited as such in the caption and are never presented as documentary photographs of a specific event.
Accountability
A named human editor is responsible for every published article, including those drafted with AI assistance. Our editorial standards describe the verification stages a draft must clear, and our corrections policy covers what happens when we get something wrong.