Automation and AI Policy

Where Registry Recognizer uses automation, what it cannot decide, and how human review works.

How automation is used

Software assists with collecting public source material, normalizing fields, linking related records, detecting duplicates, generating maps and indexes, and preparing summaries or article drafts. Automated tools can make mistakes, omit context, or preserve an error from a source.

Automatically compiled public records

Booking, charging, wanted, registry, and similar pages may be compiled automatically from cited public sources. They default to Public Record and Not Reviewed. They remain outside the Google News sitemap, News RSS feeds, and NewsArticle markup.

Human-approved reporting

Automation cannot approve an article for News. A human editor must review the article, choose an accurate non-public-record content type, confirm its source attribution and presentation, and explicitly mark it Approved. The server then records the responsible desk and review time.

Disclosure

When automation or AI materially contributes to a News article, the article carries a visible production note. The disclosure travels with the canonical article metadata so it appears on the public page. Editorial approval does not erase the disclosure.

Reporting a problem

Send suspected automated errors through the Corrections process and include the page URL and authoritative supporting source.

Last reviewed July 22, 2026.