We use AI tools to help draft from public records. Every story is edited by Peter Gorman before it publishes. Desk pen names are beat labels, not staff reporters.

The Talbot Patriot is produced with AI assistance. Public records, agency releases, and published local reporting are the source material. The voice, the headlines, and the judgment about what belongs on a Talbot front page are editorial choices, not a wire-service dump.

Stories are drafted with AI from named public records and published reporting. Peter Gorman, editor, edits and decides what publishes. Bylines such as Helen Voss or Priya Lang are desk names for a beat (Civic Affairs, Schools), not separate employees. If a fact is wrong, the paper — not the desk name — corrects it.

Human review

No story publishes without a person reading it first. Peter Gorman reads every article, checks numbers against the source document, and kills what does not hold. Drafting is automated. Publishing is not.

What can go wrong

AI drafting invents things that do not look invented — a quote, a figure, a name attached to the wrong vote. That is a known failure of the tool, not an occasional slip. It is why a person reads every story, and why every number here should trace to a named agency document. If it does not trace, it does not run.

Corrections

Tell us through Contact. Errors of fact are corrected within 48 hours, and the story carries a dated note saying what was wrong and what changed. We do not quietly edit and move on. If the error came from the drafting tool rather than the source, the note says that.

Opinion and elections

Desk names do not hold opinions. No AI-assisted byline takes a position on a ballot question, a candidate, or a pending vote of the County Council, a town commission, or the Board of Education. Opinion and Letters runs under real names only. Stories laying out competing positions run as Analysis and say so.

This is not a substitute for a full-time courthouse reporter sitting in every room. Meeting recaps are based on posted agendas, minutes, and contemporaneous accounts. Where a number is used (MCAP scores, budgets, sanctuary acres), it is taken from the named agency document. If we get one wrong, say so through Contact and we will correct it in the story.

Opinion & Letters is labeled. Analysis is labeled. Comments on stories are signed notes, held for review — they are not a poll. The rules sit on Contact.

We will not use generated photographs of real people as if they were news pictures. Article images are illustrations. Officials, votes, and quotes come from the public record.