Named attribution
Each citation must identify a named speaker — a person, not an anonymous source or a corporate entity speaking in the passive voice.
How citations.press selects, verifies, and publishes source-backed facts — and the journalists whose work on AI, trust, and citation informs our editorial direction.
citations.press is a citation index, not a news wire or an opinion outlet. We extract quantified facts from named publications and from brand-submitted press releases, structure them for discovery, and publish only what passes human editorial review. Every indexed fact is tied to a named person and a number, drawn verbatim from the source, and linked back to the original article.
We built this index for a moment when AI search and answer engines are reshaping what counts as a credible source. Machines need structured, attributable facts; journalists need provenance they can trust; readers need to know where a claim came from. citations.press sits in that gap — indexing what was said, by whom, with what number, and where it was first reported.
We do not guarantee that any AI system will cite us. We do not present ourselves as the original publisher of third-party facts.
Every citation enters the queue as pending. Nothing appears on the public site or in /citations.json until an editor approves it. Our selection criteria are deliberately narrow — quality over volume. See also our source selection policy.
Each citation must identify a named speaker — a person, not an anonymous source or a corporate entity speaking in the passive voice.
Each citation must include at least one numeric metric: a count, percentage, dollar figure, date-bound statistic, or comparable measurable fact.
The supporting quote is taken directly from the source article or press release — not paraphrased, not summarised by a model.
Every citation links to the original publication. citations.press attributes facts to their source; we never position ourselves as the claimant.
Automated extraction assists discovery, but approval is always editorial. Editors verify attribution, check the excerpt against the source, and reject entries that fail our criteria.
Press releases submitted by brands are clearly labelled Brand submission on the site. The trust model depends on this transparency.
The questions behind citations.press — what AI should cite, what counts as credible, how journalism survives automated summarisation — are being debated by the reporters who cover them every day. The journalists below represent the editorial perspectives we are building around.
They have not been interviewed for this site. Their inclusion reflects the themes we intend to explore and the authority we want this index to earn — not endorsements or published quotes.
Journalist and author
Chris Stokel-Walker is a UK-based journalist and author specialising in technology, artificial intelligence, digital culture, and online platforms. His work has appeared in major outlets including The Times, The Sunday Times, The Economist, Bloomberg, the BBC, WIRED, New Scientist, and The Guardian. He has written extensively about how technology changes media, creators, search behaviour, and public trust.
“How AI search is changing what counts as a credible source.”
A topic we aim to explore — not a quote from Chris.
AI journalist
Karen Hao is an award-winning journalist covering the social impact of artificial intelligence. She co-hosts the BBC podcast The Interface, contributes to outlets including The Atlantic and More Perfect Union, and co-created the Pulitzer Center's AI Spotlight Series, which trains journalists around the world on how to cover AI.
“What AI systems should and should not cite.”
A topic we aim to explore — not a quote from Karen.
If you believe an indexed citation misrepresents its source, please see our corrections policy or request a correction. For editorial tips or partnership enquiries, use our contact form.
Read our full Terms & Conditions for acceptable use, attribution requirements, and submission policies.