{"paper":{"title":"Who Gets Flagged? The Pluralistic Evaluation Gap in AI Content Watermarking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"AI content watermarking produces unequal detection rates across languages, cultures, and demographic groups because its signals depend on varying content statistics.","cross_cats":["cs.CL","cs.CR","cs.CV"],"primary_cat":"cs.CY","authors_text":"Alexander Nemecek, Erman Ayday, Osama Zafar, Wenbiao Li, Yuqiao Xu","submitted_at":"2026-04-15T12:06:56Z","abstract_excerpt":"Watermarking is becoming the default mechanism for AI content authentication, with governance policies and frameworks referencing it as infrastructure for content provenance. Yet across text, image, and audio modalities, watermark signal strength, detectability, and robustness depend on statistical properties of the content itself, properties that vary systematically across languages, cultural visual traditions, and demographic groups. We examine how this content dependence creates modality-specific pathways to bias. Reviewing the major watermarking benchmarks across modalities, we find that, "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"watermarking is held to a lower fairness standard than the generative systems it is meant to govern.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That watermark signal strength, detectability, and robustness depend on statistical properties of the content that vary systematically across languages, cultural visual traditions, and demographic groups.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AI content watermarking exhibits detection disparities across languages, cultures, and demographics due to content-dependent signal properties, with benchmarks failing to disaggregate performance and watermarking held to lower fairness standards than generative models.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AI content watermarking produces unequal detection rates across languages, cultures, and demographic groups because its signals depend on varying content statistics.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"677513176822051cbc0ee28615d1f5df711eaab73965ebbf09159ffaa999b654"},"source":{"id":"2604.13776","kind":"arxiv","version":2},"verdict":{"id":"a1459f03-80df-45d4-b162-f178494766ce","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T12:24:09.451187Z","strongest_claim":"watermarking is held to a lower fairness standard than the generative systems it is meant to govern.","one_line_summary":"AI content watermarking exhibits detection disparities across languages, cultures, and demographics due to content-dependent signal properties, with benchmarks failing to disaggregate performance and watermarking held to lower fairness standards than generative models.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That watermark signal strength, detectability, and robustness depend on statistical properties of the content that vary systematically across languages, cultural visual traditions, and demographic groups.","pith_extraction_headline":"AI content watermarking produces unequal detection rates across languages, cultures, and demographic groups because its signals depend on varying content statistics."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.13776/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}