As generative AI drives down the cost of polished prose, fluency can no longer serve as a reliable proxy for truth. We term this failure mode the Fluency Trap: users tend to trust fluent hallucinations while discounting accurate information once it is disclosed as AI‑generated. Binary “Made with AI” labels only reveal authorship and do not show what evidence backs each claim. To address this, we introduce Provenance Density—an evidence‑visualization interface that displays the density of verified claims throughout a text.
In a user study with 81 participants, an idealized Provenance Density interface created a large discernment gap between truth and fabrication ($+4.15$ points, $d=1.82$), whereas participants receiving no signal showed no detectable discrimination. A technical audit of 200 samples revealed that retrieval density alone is insufficient; unexpectedly, the Consistency Veto carries most of the discriminative signal on dynamic queries.
As AI‑generated content becomes indistinguishable from human writing, effective transparency must shift from mere authorship disclosure toward evidence visualization.
Review: Provenance Density offers a concrete, user‑friendly way to surface the credibility of information in real time, making it a promising candidate for the next generation of transparency tools in an AI‑saturated information landscape.