How we use AI

Last updated: October 24, 2026

Brandon Dispatch is a small newsroom. We use software, including AI, to cover more public records than a small team can read by hand. This page tells you where we use AI and where a person checks the work.

Where we use AI

Finding records. Our software reads public sources every day: county hearings, state restaurant inspections, the Sheriff's Office, the National Weather Service, road closures, library and county calendars and more.
Writing drafts. An AI model (Claude, made by Anthropic) writes a first draft of each story from the record. Our rules tell it to use only facts that are in the record, and to link the record in the story.
Short text around a story. AI also writes search titles, newsletter summaries, social media posts and the questions we ask readers.
Summaries of other news outlets. Sometimes we write a short summary of a story from another news outlet. We name the outlet and link to its story.
Daily video. AI writes the script for our daily video. The voice in the video is an AI copy of our publisher's voice, made with ElevenLabs with the publisher's permission. We mark these videos as AI-generated on Facebook and Instagram.

Where a person checks the work

A person reviews every story that involves people, crime or safety before it goes live.
Routine stories can publish without a person reading them first. Examples are road closures, restaurant openings and event lists. Before one of these stories goes live, our software checks the names, numbers and dates in the draft against the record. If a check fails, the story goes to a person.
A person reviews every story that starts from a reader's tip.

What we do not do

We do not use AI to make or change news photos.
We do not publish quotes that AI made up. A quote in a story comes from a record or from a person who said it.
We do not let sponsors or AI decide what we cover.

Your data

When we write a story from a reader's tip, the AI service gets the text of the tip. It does not get your name or email address. Under its commercial terms, Anthropic does not use this data to train its models. Read our privacy policy.

Mistakes

AI can make mistakes, and so can people. If you see an error, email correction@brandondispatch.com. Read our corrections policy.