<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhao-etal-2026-provenance">
    <titleInfo>
        <title>The Provenance Horizon: Short Regulated Text Resists Reliable Watermarking Under Editing</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yanmin</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jikun</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Tengjiao</namePart>
        <namePart type="family">Zheng</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yang</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dongxin</namePart>
        <namePart type="family">Guo</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">S</namePart>
        <namePart type="given">M</namePart>
        <namePart type="family">Yiu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>The EU Artificial Intelligence Act now requires providers to mark generated text so its origin can be checked. In regulated industries the duty is sharper: a marker must never flag third-party writing, survive routine review edits, and work on outputs often a sentence long. We call these the regulated-provenance triad and prove that no bounded-distortion watermark perturbing a model’s token distribution can satisfy all three below a critical length, the provenance horizon. The proof combines the Le Cam two-point method with a strong data-processing inequality for the edit channel; a matching achievability bound for the green-list family under a memoryless resampling channel makes the scaling order-tight. The barrier rises at negligible-distortion budgets, not at deployed ones. The contraction factor entering the bound is a proven upper bound for the memoryless resampling channel and a calibrated estimate for realistic paraphrase, under which we estimate most sentence-length regulated outputs fall below the horizon. Relocating provenance to a tamper-evident log read through a conformally calibrated retrieval test sidesteps the barrier: under exchangeable calibration its false-positive rate is at or below the target at every length. Coverage carries no guarantee and degrades under an adaptive editor.</abstract>
    <identifier type="citekey">zhao-etal-2026-provenance</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.1813/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>41316</start>
            <end>41333</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T The Provenance Horizon: Short Regulated Text Resists Reliable Watermarking Under Editing
%A Zhao, Yanmin
%A Wu, Jikun
%A Zheng, Tengjiao
%A Liu, Yang
%A Guo, Dongxin
%A Yiu, S. M.
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhao-etal-2026-provenance
%X The EU Artificial Intelligence Act now requires providers to mark generated text so its origin can be checked. In regulated industries the duty is sharper: a marker must never flag third-party writing, survive routine review edits, and work on outputs often a sentence long. We call these the regulated-provenance triad and prove that no bounded-distortion watermark perturbing a model’s token distribution can satisfy all three below a critical length, the provenance horizon. The proof combines the Le Cam two-point method with a strong data-processing inequality for the edit channel; a matching achievability bound for the green-list family under a memoryless resampling channel makes the scaling order-tight. The barrier rises at negligible-distortion budgets, not at deployed ones. The contraction factor entering the bound is a proven upper bound for the memoryless resampling channel and a calibrated estimate for realistic paraphrase, under which we estimate most sentence-length regulated outputs fall below the horizon. Relocating provenance to a tamper-evident log read through a conformally calibrated retrieval test sidesteps the barrier: under exchangeable calibration its false-positive rate is at or below the target at every length. Coverage carries no guarantee and degrades under an adaptive editor.
%U https://aclanthology.org/2026.emnlp-main.1813/
%P 41316-41333

Markdown (Informal)

[The Provenance Horizon: Short Regulated Text Resists Reliable Watermarking Under Editing](https://aclanthology.org/2026.emnlp-main.1813/) (Zhao et al., EMNLP 2026)

• The Provenance Horizon: Short Regulated Text Resists Reliable Watermarking Under Editing (Zhao et al., EMNLP 2026)

ACL

• Yanmin Zhao, Jikun Wu, Tengjiao Zheng, Yang Liu, Dongxin Guo, and SM Yiu. 2026. The Provenance Horizon: Short Regulated Text Resists Reliable Watermarking Under Editing. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 41316–41333, Budapest, Hungary. Association for Computational Linguistics.