Full Disclosure, Less Trust? How the Level of Detail about AI Use in News Writing Affects Readers’ Trust

Author: Pooja Prajod, Hannes Cools, Thomas Roeggla, Karthikeya Puttur Venkatraj, Amber Kusters, Alia ElKattan, Pablo Cesar, Abdallah El Ali · Type: academic · Status: draft · Published: 2026-01-14 · URL: https://arxiv.org/abs/2601.09620

Licence: arXiv preprint; check arXiv license on the abstract page before reuse (typically arXiv non-exclusive license or CC).

Summary

Experimental study on AI-use disclosure in news. Detailed disclosures reduced reader trust while both one-line and detailed disclosures increased source-checking; about two-thirds of participants nonetheless preferred comprehensive disclosure, and minimal-disclosure fans wanted detail-on-demand formats. Frames disclosure as a transparency-versus-trust trade-off rather than a dilemma.

Insights

• Detailed AI-use disclosure measurably lowers trust in the disclosed content even when readers judge the content accurate - transparency has a real cost that must be designed around, not ignored. (high, draft) disclosure trust content quality

  Trust declined specifically when detailed AI disclosures were provided, while source-checking behavior increased for both one-line and detailed disclosures. — Key findings / abstract

• Readers still want disclosure despite the trust penalty, and a detail-on-demand format (brief label with expandable process detail) is the design most aligned with stated preferences. (high, draft) disclosure trust marketing

  Approximately two-thirds of participants favored comprehensive disclosure information; those preferring minimal disclosure expressed a need for detail-on-demand disclosure formats. — Participant preferences

• Disclosure outcomes are a manageable trade-off, not a binary dilemma; how much detail is shown, and how, changes the trust impact. (medium, draft) disclosure trust

  Not all AI disclosures lead to a transparency dilemma, but instead reflect a trade-off between readers’ desire for more transparency and their trust in AI-assisted news content. — Conclusions