AI policy

[DRAFT: first-draft copy awaiting the owner's sign-off. Not final.]

AutoMotor News uses AI to help prepare its stories. A person reviews every item before it is published. This page explains what each does.

What AI does

We use AI models (Claude, made by Anthropic) to:

• watch public sources, such as new recall filings and court and government releases, and flag possible stories;
• draft stories, headlines and newsletter issues from those sources and from our own data;
• fact-check each draft, claim by claim, against the sources it cites.

Figures in our data studies are computed by our own code, not written by an AI model.

What people do

A qualified person reviews every item before publication: stories, headlines, updates, corrections, newsletter issues and site pages. They can approve it, send it back with changes or reject it. Nothing is published without that approval, and the reviewer's name appears in the byline: "AutoMotor News Newsroom, reviewed by" followed by their name. Reviewers are listed on our Newsroom page.

Where a page needs first-hand input, such as notes on what surprised us in a data study, a person writes it.

How we disclose it

Every story carries a note saying it was prepared with AI assistance, with a link to this page. So do our newsletter and the emails we send. We don't add made-up human authors.

We never disguise AI use

Text produced by AI tools may carry watermarks or other signals that show it was machine-generated. We leave them alone. We don't use "humanizer" or paraphrasing tools, character tricks or any other method meant to hide AI involvement or get around watermarks or AI detectors.

What we never do

• Invent people, quotes, reviews, testimonials, experts or statistics.
• Use AI-generated photos to show real events.
• Publish anything a person has not reviewed.

Mistakes

AI tools and people both make mistakes. If you find one, write to corrections@automotor.news. Our corrections policy explains what happens next.