Standards

Last updated: September 14, 2026

Most articles on this site were drafted by software, directed by or overseen by a person. This page explains exactly what that means.

First, we do not invent staff writers. There is no such thing on this site as a byline belonging to a person who does not exist, and no desk byline carries a photograph of a face, because a face asserts a person.

What the software does

• Drafts from verified source data. Our pipeline gathers public records, official data feeds, and published reporting, then drafts an article from that material. It is not asked to recall facts from memory; it is given sources and asked to write from them. We do not pretend to report on anything, merely package information and data from other sources.
• Checks its own claims against those sources. Every factual claim is traced back to the source material it came from. A draft where too many claims can't be traced does not publish.
• Checks the risky claims against independent sources. Claims that could hurt someone if wrong — a person named in connection with wrongdoing, a death, an arrest, a business closing, an official vote, a dollar figure, a scheduled date — are re-checked against live sources that are not the original. A claim that independent sources contradict fails outright. A claim nobody can corroborate is held for a person to look at.

Where the sports numbers come from

Scores, schedules, standings and statistics on our sports pages come from ESPN's public site API and the MLB Stats API; college football ratings and lines from CollegeFootballData; forecasts from the National Weather Service. Those are the facts, and every page that shows them says so in its footer. Everything written on top of them is ours: play descriptions are composed by our software from each play's recorded fields, never copied from a feed's text; the win-probability figures and charts are our own model, labeled as such, and are rougher than a broadcaster's; previews, recaps, rankings and analysis are the desk's work from those records. We do not show player photographs we do not have the rights to, and we do not read other publishers' articles to write our analysis — a headline, a one-line summary and our own records are what the writer is given.

What a person does

A human editor guides the system — rules for factchecking, setting tone and themes in content, and determining frequency of publishing and distribution. Editors can send a draft back, or publish one the automated checks flagged, but only after human review.

What we don't do

• We don't publish articles attributed to people who don't exist.
• We don't use synthetic photographs of people, ever.
• We don't run AI illustrations on hard news or sports. If we were ever to use a generated image — an editorial column, a recipe, a video background — it is credited "Illustration: AI-generated" on the page, and the image file itself carries the IPTC digital source type that marks it as synthetic, so the disclosure travels with the file even off our site.
• We don't rewrite another newsroom's exclusive and call it ours. A story resting on a single outlet's original reporting is held for review and, if it runs, runs as a brief that links to them.
• We also do not link to third-party news immediately to give those outlets time to break news on their own terms.

The same story across our sites

We publish local brands in many towns off one newsroom. When a regional or national story matters in several of them, only one version is submitted to search engines; the copies on our other sites are marked noindex and left out of those sites' sitemaps. Where we rewrite a story substantially for a particular town, that version stands on its own. We measure this: each site tracks what share of its pages are its own work, and a site that falls below our floor doesn't get to expand until it comes back up.

Corrections

If something here is wrong, tell the editor of record named on the article, or write to corrections@commonwires.com. We correct on the article itself and say what changed. An article that turns out to be wrong at its foundation is pulled, not quietly edited.

Related: About · Privacy Policy · Terms of Service