Beyond 'Made with AI': Provenance Density Visualizes Verified Claims to Counter the Fluency Trap

Forum topic · 小凯 · 2026-09-06

Summary

This arXiv paper (2509.00010) by Qing Zhang, Yifei Huang, and Juyoung Lee introduces Provenance Density, an evidence-visualization interface designed to address what the authors call the 'Fluency Trap': users trusting fluent AI hallucinations while discounting accurate content once labeled as AI-generated. Instead of binary 'Made with AI' authorship labels, the interface displays the density of verified claims within a text. In a user study with 81 participants, an idealized Provenance Density interface produced a large discernment gap between true and fabricated content (+4.15 points, d=1.82), while participants without any signal showed no detectable discrimination. A technical audit of 200 samples found that retrieval density alone is insufficient; surprisingly, a 'Consistency Veto' mechanism carried most of the discernment signal on dynamic queries. The paper argues that effective transparency should shift from authorship disclosure toward evidence visualization as AI-generated content becomes indistinguishable from human writing.

Paper Overview

Field: AI/ML Authors: Qing Zhang, Yifei Huang, Juyoung Lee Published: 2026-09-06 arXiv: 2509.00010

Abstract

As generative AI makes polished prose cheap to produce, users can no longer rely on fluency as a proxy for truth. We call this failure mode the Fluency Trap: users trust fluent hallucinations while also discounting accurate content once it is disclosed as AI-generated. Binary 'Made with AI' labels respond with authorship disclosure, but they do not show what supports a claim. We propose Provenance Density, an evidence-visualization interface that shows the density of verified claims in a text. In a user study with 81 participants, an idealized Provenance Density interface produced a large discernment gap between truth and fabrication (+4.15 points, d=1.82), whereas participants given no signal showed no detectable discrimination. A technical audit with 200 samples shows that retrieval density alone is insufficient; unexpectedly, a Consistency Veto carries most of the discernment signal on dynamic queries. As AI-generated content becomes indistinguishable from human writing, effective transparency must shift from authorship disclosure to evidence visualization.

Key Findings

The Fluency Trap: Users trust fluent hallucinations, while accurate AI-generated content gets discounted once its authorship is disclosed — fluency and labels both fail as truth signals.
Provenance Density interface: Instead of binary 'Made with AI' labels, the proposed UI visualizes the density of verified claims in a text.
User study (n=81): The idealized interface created a large discernment gap between truth and fabrication (+4.15 points, d=1.82); participants with no signal showed no detectable discrimination.
Technical audit (200 samples): Retrieval density alone is insufficient; a Consistency Veto mechanism surprisingly carried most of the discernment signal on dynamic queries.
Implication: Effective transparency should move from authorship disclosure toward evidence visualization.

--- *Auto-collected on 2026-09-06*