Humanized but Still True? Auditing Facts and Detection in AI-Text Humanizers
Factual evidence
What the source reports
Research audits AI-text humanizers, finding rewrites often corrupt factual details like numbers, dates, and names to evade detection.
Inspect the evidence
- Inclusion basis
- Enterprise AI
- Publisher and source type
- arXiv cs.CL — Computation and Language · RESEARCH
- Published by source
- 4 November 2026
- Collected by OneBench
- 5 Nov 2026, 02:01 UK
Stored source excerpt
arXiv:2611.00359v1 Announce Type: new Abstract: AI "humanizers" rewrite machine-generated text so that detectors classify it as human-written. Public leaderboards score evasion and embedding similarity, but…
Short excerpt from the collected text, not the full source. Use the source link to read it in context.
The factual summary is a OneBench synthesis, not a quotation or independent verification. Collection time is not publication time. Open the source for its full context; related reporting can share the same underlying announcement.
OneBench interpretation
Institutional assessment
So what
Tools used to obscure synthetic text alter critical factual entities, introducing undetected hallucinations into automated documentation workflows.
Do what
Review model governance controls regarding third-party text humanizers or unverified synthetic text generation in operational reporting.