Editorial policy and use of AI
This policy explains how Plainwire chooses, writes, checks and corrects its stories, and how our desk uses artificial intelligence. It applies to every story, The Brief and our explainers.
Our principles
- Plain and factual. Short sentences, the key fact first, no exclamation marks, no teaser headlines. Headlines are in sentence case and no longer than 75 characters.
- Sourced. Every story names the outlet whose report it is based on and links to that report.
- Checked by people. No story is published until an editor on our desk has approved it.
- Corrected openly. When we get something wrong, we fix it and say so.
How we choose stories
We favor stories that many outlets are reporting, that are fresh, and that matter to readers in the US, the UK, Canada, Australia and Ireland. We aim for a balance between sections every day. We do not publish live blogs, opinion columns, sponsored or partner content, horoscopes, lottery results, TV guides, quizzes, shopping deals, sport, celebrity gossip, gaming or cryptocurrency news. We do not publish the same story twice.
How we use AI
Plainwire uses AI tools as a writing aid. For each story, an AI model writes a version in plain English of a report published by a named outlet: in its own words when the report is in English, or as an English version when it is in another language. An AI model may also suggest the topics a story belongs to.
Before publication, an editor on our desk:
- reads the version against the original report;
- checks names, dates, numbers and quotes, and removes anything the source does not support;
- edits the headline and the standfirst, which answers who, what, when and where;
- writes or fixes the “Why it matters” box (two to four sentences of context);
- rejects the story if it cannot be checked or does not meet this policy.
Our system automatically holds any version in which a number does not appear in the source; such a story can only be published after an editor has checked the figures. Every story says, in its source box, that it is an AI-assisted version checked by our desk.
Verification
We rely on established news outlets and say which one each story comes from. When outlets disagree, we say so or we wait. We do not add facts that are not in the source, and we attribute claims to the people who made them.
Images
We never show the original publisher’s photos. Our pictures come from openly licensed collections such as Wikimedia Commons, from Pexels, from official press kits or from news agencies, always with their credit and, when the license requires it, a link to the license. When no suitable picture exists, we use our own artwork, labeled “Illustration”.
Corrections and takedowns
If we make a mistake, we correct the story and add a dated note under it explaining what changed. All notes are listed on our corrections page. Simple typos are fixed without a note. If a story cannot be corrected, or if a rights holder or a person concerned makes a justified request, we remove it: its address then answers that the story has been removed. To report an error or request a takedown, see our contact page.
Independence and ethics
Plainwire is not affiliated with any party, government or company it covers. Advertising, once switched on, is clearly labeled and never influences our choice of stories. We do not accept payment for coverage. Stories are signed by the editor who approved them, or by the Plainwire desk; we never invent authors.
Bylines
Each named editor has an author page with their role and how to contact them. Stories signed “Plainwire desk” are approved by a member of the desk.