From Forensics to Ecosystems: Rethinking Watermarks for Generative AI Oversight

Daniel Susser; Gili Vidan; John Thickstun

REVIEW 4 major objections 3 minor 69 references

From Forensics to Ecosystems: Rethinking Watermarks for Generative AI Oversight

T0 review · 4 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper argues that AI watermarks should be judged as ecosystem measurement tools, not as forensic authenticators, and that their statistical weakness is a feature when measuring aggregate synthetic content.

desk verdict A genuinely useful reframing of watermarks as ecosystem metrics, but the 'more tractable' claim rests on unsupported assumptions about attack costs and provider coverage. read the letter →

arxiv 2608.07337 v1 pith:TGUGL3NW submitted 2026-08-07 cs.CY

Daniel Susser , John Thickstun , Gili Vidan This is my paper · ORCID

classification cs.CY

keywords AIwatermarkingsyntheticcontentecosystemmeasurementdigitalforensicsgenerativegovernancestatisticaldetectionprovenancemediasaturation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that the standard critique of AI watermarks—that they are brittle, ambiguous, and easy to evade—rests on a forensic assumption that each watermark must authenticate an individual piece of content. The authors propose an "ecosystems approach": watermarks are best used to measure the overall saturation of synthetic content in a media environment, like wastewater testing tracks disease at the community level. On this view, a watermark need not survive every adversarial attempt at removal; it only needs to remain detectable often enough that aggregate statistics stay meaningful. The paper contends that this reframing makes the governance challenges of watermarking more tractable, because it lowers the technical bar for watermark strength and aligns incentives for adoption by centralized AI providers. A sympathetic reader should care because it rescues watermarking as a viable policy tool even where forensic identification is hopeless.

What carries the argument

The conceptual engine is the "ecosystems approach", defined as the use of watermarks as statistical indicators of overall synthetic content prevalence in a given information ecosystem, rather than as forensic authenticators of individual content. Operationally, it relies on the statistical nature of AI watermark detection, in which a detector scans content and returns a p-value measuring the probability of observing that content under the null hypothesis that it was generated without a watermark. The approach also depends on provider centralization: watermarking at the moment of synthesis by a few large model providers can imprint a critical mass of synthetic content with a detectable signal. Supporting machinery includes the three watermark desiderata—detectability, quality preservation, and removal friction—and the entropy constraint that makes text watermarks weak relative to images or video, which the ecosystem view tolerates.

What would settle it

A concrete test: on a major text-based platform, measure the fraction of all synthetic content that carries a detectable watermark by comparing a representative sample against a comprehensive ground-truth audit; if adversarial scrubbing or unwatermarked self-hosted model output means that most synthetic content escapes watermark detection, then the aggregate signal would not track true synthetic content saturation.

Watch

Extended reading notes

Core claim

The central claim is that digital watermarks for generative AI should be reconceptualized from forensic evidence about individual pieces of content into population-level indicators of synthetic content saturation. The authors argue that critics are right that watermarks cannot reliably answer "is this AI?" for a given text, image, or recording, but that this failure is irrelevant to the ecosystems question "how much AI is all around us?". Because detection is inherently statistical, returning a p-value rather than a certainty, watermarks are naturally suited to aggregate measurement; and because a handful of major AI providers dominate the market, watermarks applied by just a few providers could cover a significant fraction of all synthetic content. The paper develops this position through a technical overview of watermarking, an analysis of five governance questions reframed by the ecosystem perspective, and two concrete scenarios—music streaming and scientific publishing—showing how aggregate watermark signals could inform platform policy and institutional oversight.

Load-bearing premise

The argument depends on the premise that although individual watermarks are brittle, removal and evasion are hard enough at scale that aggregate statistics remain reliable, and that enough synthetic content is watermarked—through provider centralization and cooperation—for those statistics to be meaningful.

Editorial extensions

If this is right

Watermarks become tools for system-level intervention, such as reducing aggregate synthetic content on a platform, rather than for punishing individual authors or posters.
Weak watermarks, including those in low-entropy text, remain useful because aggregate statistics can be compiled from many individually weak signals.
The detection dilemma loses much of its force: even if public detectors allow adversaries to learn circumvention strategies, ecosystem monitoring can still function with representative sampling and aggregate statistics.
Major AI model providers, facing incentives to support monitoring rather than individualized sanctions, may more readily adopt watermarking, yielding widespread coverage of synthetic content.
Platforms and institutions could publish aggregate saturation statistics—for example, the fraction of music streams or submitted papers carrying a synthetic signal—as a form of transparency and friction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the ecosystem approach is adopted, the design goal for watermarks shifts from maximizing individual robustness to maximizing the reliability of aggregate estimates, which could lead to different trade-offs in watermark strength and public detection access.
Regulators could treat aggregate watermark statistics as a standardized disclosure metric, analogous to transparency reports, raising new questions about sampling methodology and false-positive bias at the population level.
The framework implies that watermark keys need not be public for aggregate measurement to work, so providers could keep detection private while still enabling trusted third-party audits with access to samples.
A testable extension: platforms could run a controlled comparison of watermarked synthetic content with and without adversarial stripping to measure how much scrubbing is needed before aggregate saturation estimates become meaningfully biased.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper's core move—shifting watermarks from forensic identification of individual synthetic artifacts to aggregate saturation metrics for information ecosystems—is genuinely useful and well-argued. The wastewater analogy is apt: individual testing is unreliable, but population-level signals can still inform policy. The paper is honest about risks (surveillance, complacency, forgery, biased false positives) and engages the critical literature (Zhang et al. 2024; Sadasivan et al. 2024; Fernandez et al. 2024). It's a position paper; the absence of new empirical results is structural, not a flaw.

The central tractability claim is the soft spot. The argument that brittleness doesn't matter at scale depends on two premises the paper asserts but doesn't defend: that stripping watermarks is 'complicated, even for skilled attackers, at scale,' and that 'a few major providers' watermark 'a significant fraction of all synthetic content.' The first runs against Sadasivan et al. 2024, showing paraphrase-based removal of text watermarks is cheap and scalable; the paper cites that work but doesn't engage with it, just asserts the opposite. The second ignores open-weight and self-hosted models, which can generate large volumes of unwatermarked content precisely in the spam/astroturfing settings motivating oversight. If evaders strip watermarks from targeted classes, the aggregate saturation estimate is biased, not merely imprecise—attenuated by the product of coverage and evasion rates. The paper's own Risks section concedes adversarial scrubbing but asserts without evidence it's less likely under monitoring than forensics. The reframing relocates the brittleness problem rather than solving it.

That said, I'm not dismissing it. The reframing is valuable even if the tractability claim is shaky: a biased saturation index could still be informative if the bias direction and magnitude were studied. The scenarios (music streaming, scientific publishing) ground the abstraction well, and the writing is clear.

This deserves serious peer review. I'd be skeptical on the 'more tractable' claim, but the paper is a good reading-group discussion and worth citing for anyone working on AI governance.

Referee Report

4 major / 3 minor

Summary. The paper argues that digital watermarking for generative AI has been assessed primarily as a forensic instrument for identifying individual pieces of synthetic content. The authors propose an 'ecosystems approach' in which watermarks are used as aggregate statistical indicators of the saturation of synthetic content in a media environment, rather than as authenticators of particular outputs. After reviewing watermarking techniques and their desiderata, the paper reframes robustness concerns, evidence interpretation, stakeholder incentives, and governance risks under this ecosystem framing. It then illustrates the proposal with two scenarios, a music streaming service and scientific preprint publishing, and concludes that ecosystem-level use is more tractable than forensic use, provided a critical mass of synthetic content is watermarked.

Significance. If the argument succeeds, the paper is a valuable conceptual contribution to AI governance debates: it articulates a clear distinction between forensic and ecosystem uses of watermarks, and it shows how the statistical, probabilistic nature of watermarks fits aggregate measurement better than individual authentication. The paper is careful and honest, explicitly identifying risks such as surveillance via personalized watermarks, complacency, forgery, and biased false positives, and it does not oversell watermarks as a complete solution. It also grounds the discussion in concrete technical background and policy context. The main weakness is that the central empirical premises, that scaled evasion is hard and that few-provider coverage yields a representative sample, are asserted rather than demonstrated, and some of the cited literature cuts against these premises. Since these premises are load-bearing, the paper needs to engage with them directly or qualify the scope of its central claim.

major comments (4)

[Watermarks as Friction] The claim that 'it would be complicated, even for skilled attackers, to disable watermarks at scale' is load-bearing for the ecosystem approach, but the paper does not support it, and the literature it cites suggests the opposite for text. Sadasivan et al. (2024) show that paraphrase-based attacks remove text watermarks with modest resources, and Zhang et al. (2024) provide impossibility results for strong watermarking. If paraphrase attacks are cheap and automatable, scaled evasion does not require per-content labor, so aggregate saturation estimates would be attenuated by the evasion rate. The paper should either engage these results directly and state the conditions under which scaled evasion is genuinely hard, or weaken the claim to a narrower scope, for example high-bandwidth media or watermarking protocols that are specifically robust to paraphrasing.
[Conclusion and Risks of Watermarks] The conclusion's premise that 'convincing just a few major providers to implement AI watermarking would result in watermarks on a significant fraction of all synthetic content' is unquantified and unproven, and the paper does not explain how a saturation statistic would be estimated from watermark p-values. Open-weight, self-hosted, and non-cooperating providers can generate large volumes of unwatermarked text, particularly in spam, astroturfing, and manipulation settings, which are arguably the most policy-relevant cases. The 'Risks of Watermarks' section itself acknowledges blind spots from bespoke providers and scrubbed content. Without a coverage estimate or a calibration strategy, the ecosystem metric measures watermarked-provider saturation, not synthetic-content saturation. The paper should state the coverage assumptions and discuss how the metric could be validated, for example by comparison with independent estimators.
[Who are Watermarks For? / Scenarios] The paper's central comparative claim, that ecosystem-mode governance challenges are 'more tractable' than forensic ones, is asserted rather than demonstrated. The 'Who are Watermarks For?' section identifies access to a 'sufficiently representative sample of content' as the central access problem, and the music-streaming scenario concedes that platform cooperation may not align with the platform's incentives. These are not obviously easier problems than forensic interpretation: biased sampling can be as damaging to an aggregate estimate as misclassification is to individual identification. The paper should provide a comparative analysis of the two modes' failure modes, or at least a more explicit account of why the institutional demands of ecosystem measurement are lighter than the interpretive demands of forensic measurement.
[Risks of Watermarks] The paper asserts that 'adversarial scrubbing is more likely in scenarios where watermarks are deployed for forensics or moderation ... than in contexts where they are used for monitoring and ecosystem analysis,' but this claim is not argued. If aggregate statistics are used to justify stricter platform policies, content producers who benefit from synthetic content have incentives to evade watermarking regardless of whether individual content is sanctioned. Because evasion can be automated, the monitoring context does not remove the threat. The paper needs to defend this incentive claim or qualify the ecosystem approach's vulnerability to scrubbing, since the validity of the aggregate measure depends on it.

minor comments (3)

[Throughout] The manuscript refers to 'Section 3.1', 'Section 3.2', and 'Section 3.5', but the text does not show numbered sections; the cross-references should be made consistent with the actual section labels.
[Scientific Writing scenario] The arXiv example is presented as an instance of ecosystem analysis, but the paper states that arXiv's decision was based on heuristic assessment of the growth of synthetic submissions, not on watermark-based data. The example would be clearer if the paper explicitly labeled it as a non-watermark precedent for the kind of decision the ecosystem approach would inform.
[Technical Background, footnote 7] The contrast with wastewater testing is too quick: watermark detection may be calibrated by construction, but calibration of detection does not correct for non-coverage or selective evasion, which are the main threats to the aggregate estimate. The footnote should be revised to avoid implying that calibration solves the sampling problem.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ecosystems reframing is an interpretive proposal, and the minor self-citations are not load-bearing.

full rationale

This paper is a conceptual and governance argument rather than an empirical derivation; it contains no fitted parameters, equations, or predictions that reduce to its inputs. The central claim—that watermarks are better suited to aggregate ecosystem measurement than to forensic authentication—is an interpretive proposal supported by analogy and scenario analysis, not a quantity derived from the watermarking literature. The technical background cites external, published work (e.g., Christ, Gunn, and Zamir 2024; Zhang et al. 2024; Sadasivan et al. 2024) to establish both feasibility and brittleness, and the paper explicitly concedes adversarial scrubbing and non-coverage in its 'Risks of Watermarks' section, so those limitations are not hidden assumptions smuggled in as conclusions. The only self-citations are Kuditipudi et al. 2024 (co-authored by Thickstun) for watermark desiderata and distortion-free text watermarking, and Susser 2019 for the general point that policymakers favor disclosure regulation. Both are published, externally checkable results that do not depend on the present thesis; the ecosystem argument would stand even if those citations were replaced. The load-bearing empirical premises—that watermark removal is hard at scale and that a few major providers cover a significant fraction of synthetic content—are asserted rather than derived, which is a correctness/evidence concern, not circularity. No step in the argument assumes its conclusion, and no known result is merely renamed. Score 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on technical facts imported from the watermarking literature and on four empirical assumptions about aggregate robustness, provider coverage, representativeness, and statistical aggregation. No fitted parameters and no new entities appear.

assumptions (5)

domain assumption Watermarks can be embedded in generated content at generation time with statistical detection guarantees and minimal quality loss.
Invoked in Technical Background through Christ et al. 2024 and Kuditipudi et al. 2024; the ecosystems approach inherits this technical premise from prior work.
domain assumption Removing or evading watermarks is difficult at scale even though individual watermarks are brittle.
Stated in Watermarks as Friction: 'it would be complicated, even for skilled attackers, to disable watermarks at scale.' If false, aggregate statistics would not reflect the true volume of synthetic content.
domain assumption A critical mass of synthetic content will be watermarked because a small number of major providers dominate the market and can be induced to cooperate.
Used in the Conclusion: 'convincing just a few major providers to implement AI watermarking would result in watermarks on a significant fraction of all synthetic content.' No evidence is provided for the cooperation claim.
domain assumption Platforms or researchers can obtain a representative sample of content for ecosystem-level analysis.
Acknowledged in Who are Watermarks For: ecosystem analysis requires 'access to a sufficiently representative sample of content in the overall ecosystem.' The music scenario also notes that platform cooperation may be lacking.
standard math Weak individual watermark signals can be aggregated into statistically meaningful ecosystem-level estimates.
Assumed throughout the scenarios and in the phrase 'how much AI is all around us'; the paper does not formalize the estimator or its bias, but relies on standard statistical intuition.

how reviews work

0 comments

Cite this review

Pith. "Pith review of From Forensics to Ecosystems: Rethinking Watermarks for Generative AI Oversight." pith.science (2026). https://pith.science/paper/TGUGL3NW

@misc{pith2026260807337,
  author       = {Pith},
  title        = {Pith review of: From Forensics to Ecosystems: Rethinking Watermarks for Generative AI Oversight},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TGUGL3NW}},
  note         = {Machine review of arXiv:2608.07337}
}

read the original abstract

The arrival of generative AI as a cheap, widely accessible commercial service, and the tidal wave of AI-generated synthetic content it has unleashed, have provoked deep epistemic and social anxieties and raised difficult governance questions that policymakers are struggling to address. One approach that has attracted both enthusiasm from regulators and skepticism from researchers is digital watermarking. Signals embedded in a synthetically-generated piece of content indicating that it was AI-generated---possibly even identifying the specific systems that generated it---appear to offer a path toward mitigating risks of genAI that avoids the downsides of more interventionist strategies. But critics warn that watermarks may prove technically brittle, epistemically ambiguous, and politically ineffectual tools. In this paper, we explore the challenges and opportunities of using digital watermarking for AI governance, paying special attention to the specific problem of watermarking AI-generated text. We argue that such critiques often treat the problem of identifying synthetic content as an isolated forensic question. Instead, we propose reconceptualizing digital watermarks as tools for understanding the impacts of synthetic content on media ecosystems, rather than reliably identifying individual pieces of synthetic content. Such an ``ecosystems approach'' more effectively utilizes the features of watermarks. And while this approach raises its own governance challenges, we argue that they are more tractable than the challenges of using watermarks for digital forensics.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 62 canonical work pages

[1]

Intelligencer (New York Magazine) , year =

Max Read , title =. Intelligencer (New York Magazine) , year =

work page
[2]

Deepfakes and the Epistemic Apocalypse , volume =

Joshua Habgood. Deepfakes and the Epistemic Apocalypse , volume =. doi:10.1007/s11229-023-04097-3 , journal =

work page doi:10.1007/s11229-023-04097-3
[3]

2025 , url =

SynthID: A tool to watermark and identify AI-generated content , howpublished =. 2025 , url =

work page 2025
[4]

2023 , month = nov, day =

Executive Order 14110: Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. 2023 , month = nov, day =

work page 2023
[5]

2025 , month = jan, day =

Executive Order 14179: Removing Barriers to American Leadership in Artificial Intelligence. 2025 , month = jan, day =

work page 2025
[6]

942 (California AI Transparency Act)

Senate Bill No. 942 (California AI Transparency Act). 2024 , month = sep, day =

work page 2024
[7]

An Act regulating provenance regarding artificial intelligence (House Bill No. 90). 2025 , month = feb, day =

work page 2025
[8]

and Miller, Matt L

Cox, Ingemar J. and Miller, Matt L. , urldate =. The First 50 Years of Electronic Watermarking , volume =. 2002 , langid =. doi:10.1155/S1110865702000525 , pages =

work page doi:10.1155/s1110865702000525 2002

Show all 69 references

[9]

International Conference on Machine Learning , year=

Watermarks in the sand: Impossibility of strong watermarking for generative models , author=. International Conference on Machine Learning , year=
[10]

2023 , month = aug, day =

Leibowicz, Claire , title =. 2023 , month = aug, day =

2023
[11]

2008 , title =

Kirschenbaum, Matthew G , keywords =. 2008 , title =

2008
[12]

Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security , volume =

Chesney, Bobby and Citron, Danielle , year =. Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security , volume =. California Law Review , issn =. doi:10.15779/Z38RV0D15J , shorttitle =

doi
[13]

2024 , title =

Scheirer, Walter , keywords =. 2024 , title =

2024
[14]

2024 , month = apr, day =

Reducing Risks Posed by Synthetic Content: An Overview of Technical Approaches to Digital Content Transparency (NIST AI 100‑4). 2024 , month = apr, day =

2024
[15]

The New Yorker , year =

Immerwahr, Daniel , title =. The New Yorker , year =
[16]

2012 , title =

Blanchette, Jean-François , keywords =. 2012 , title =

2012
[17]

2025 , url =

Tian, Edward and Cui, Alexander , title =. 2025 , url =

2025
[18]

International Conference on Machine Learning , year=

Detectgpt: Zero-shot machine-generated text detection using probability curvature , author=. International Conference on Machine Learning , year=
[19]

International Conference on Machine Learning , year=

Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text , author=. International Conference on Machine Learning , year=
[20]

Transactions on Machine Learning Research , year=

Robust Distortion-free Watermarks for Language Models , author=. Transactions on Machine Learning Research , year=
[21]

Information , volume=

Digital image watermarking techniques: a review , author=. Information , volume=. 2020 , publisher=

2020
[22]

International Conference on Machine Learning , pages=

A watermark for large language models , author=. International Conference on Machine Learning , pages=. 2023 , organization=

2023
[23]

IEEE Access , volume=

A review of text watermarking: theory, methods, and applications , author=. IEEE Access , volume=. 2018 , publisher=

2018
[24]

The Thirty Seventh Annual Conference on Learning Theory , pages=

Undetectable watermarks for language models , author=. The Thirty Seventh Annual Conference on Learning Theory , pages=. 2024 , organization=

2024
[25]

arXiv preprint arXiv:2303.11156 , year =

Sadasivan, Vinu Sankar and Kumar, Aounon and Balasubramanian, Sriram and Wang, Wenxiao and Feizi, Soheil , title =. arXiv preprint arXiv:2303.11156 , year =

arXiv
[26]

, title =

Goodman, Ellen P. , title =. Nevada Law Journal , volume =
[27]

, title =

Norman, Donald A. , title =. 1998 , edition =

1998
[28]

and Gould, Sandy J

Cox, Anna L. and Gould, Sandy J. J. and Cecchinato, Marta E. and Iacovides, Ioanna and Renfree, Ian , title =. CHI'16 Extended Abstracts on Human Factors in Computing Systems (CHI EA '16) , year =. doi:10.1145/2851581.2892410 , note =
[29]

Florida Law Review , volume =

Ohm, Paul and Frankle, Jonathan , title =. Florida Law Review , volume =. 2018 , url =

2018
[30]

and Benesch, Susan , title =

Frischmann, Brett M. and Benesch, Susan , title =. Yale Journal of Law & Technology , volume =. 2023 , note =

2023
[31]

Fernandez, Pierre and Level, Anthony and Furon, Teddy , year =. What. Workshop on Generative AI and Law, at International Conference on Machine Learning , langid =
[32]

and McGregor, Sean and Ovadya, Aviv , title =

Leibowicz, Claire R. and McGregor, Sean and Ovadya, Aviv , title =. Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society (AIES '21) , year =. doi:10.1145/3461702.3462584 , note =

2021
[33]

arXiv preprint arXiv:2405.11109 , year=

Watermarking language models for many adaptive users , author=. arXiv preprint arXiv:2405.11109 , year=

arXiv
[34]

The Twelfth International Conference on Learning Representations , year=

An Unforgeable Publicly Verifiable Watermark for Large Language Models , author=. The Twelfth International Conference on Learning Representations , year=
[35]

IACR Communications in Cryptology , volume=

Publicly-Detectable Watermarking for Language Models , author=. IACR Communications in Cryptology , volume=
[36]

The 28th International Conference on Artificial Intelligence and Statistics , year=

On the Difficulty of Constructing a Robust and Publicly-Detectable Watermark , author=. The 28th International Conference on Artificial Intelligence and Statistics , year=
[37]

U.S. v. Smith Indictment (Press Release PDF)
[38]

2015 , url =

Finn Brunton , title =. 2015 , url =

2015
[39]

2019 , title =

digitalSTS , isbn =. 2019 , title =

2019
[40]

SSRN Electronic Journal , year =

Burrell, Jenna , title =. SSRN Electronic Journal , year =
[41]

Duncan and Wallach, Hanna , title =

Barocas, Solon and Guo, Anhong and Kamar, Ece and Krones, Jacquelyn and Morris, Meredith Ringel and Vaughan, Jennifer Wortman and Wadsworth, W. Duncan and Wallach, Hanna , title =. 2021 , isbn =. doi:10.1145/3461702.3462610 , booktitle =

2021
[42]

and Mitchell, Margaret and Gebru, Timnit and Hutchinson, Ben and Smith-Loud, Jamila and Theron, Daniel and Barnes, Parker , title =

Raji, Inioluwa Deborah and Smart, Andrew and White, Rebecca N. and Mitchell, Margaret and Gebru, Timnit and Hutchinson, Ben and Smith-Loud, Jamila and Theron, Daniel and Barnes, Parker , title =. 2020 , isbn =. doi:10.1145/3351095.3372873 , booktitle =

2020
[43]

2024 , month = apr, day =

Bickert, Monika , title =. 2024 , month = apr, day =

2024
[44]

, keywords =

Cole, Simon A. , keywords =. 2001 , title =

2001
[45]

American anthropologist , pages =

Goodwin, Charles , keywords =. American anthropologist , pages =. 1994 , title =

1994
[46]

1995 , title =

Jasanoff, Sheila , keywords =. 1995 , title =

1995
[47]

Regulation by Contract, Regulation by Machine , urldate =

Margaret Jane Radin , journal =. Regulation by Contract, Regulation by Machine , urldate =
[48]

2007 , title =

Gillespie, Tarleton , keywords =. 2007 , title =

2007
[49]

arXiv preprint arXiv:2503.16458 , year=

Users Favor LLM-Generated Content--Until They Know It's AI , author=. arXiv preprint arXiv:2503.16458 , year=

arXiv
[50]

2024 , month = apr, day =

Building Trust in the Age of AI , howpublished =. 2024 , month = apr, day =

2024
[51]

and Givi, Julian , title =

Kirk, Colleen P. and Givi, Julian , title =. Journal of Business Research , volume =. 2025 , doi =

2025
[52]

Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency , pages=

AI Art is Theft: Labour, Extraction, and Exploitation: Or, On the Dangers of Stochastic Pollocks , author=. Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency , pages=

2024
[53]

2014 , month = sep, day =

Rudder, Christian , title =. 2014 , month = sep, day =

2014
[54]

2011 , title =

Roberts, Dorothy , keywords =. 2011 , title =

2011
[55]

Nature , volume=

Peer review: Troubled from the start , author=. Nature , volume=. 2016 , publisher=

2016
[56]

Advancing content provenance for a safer, more transparent AI ecosystem , year =
[57]

Rickford and Dan Jurafsky and Sharad Goel , title =

Allison Koenecke and Andrew Nam and Emily Lake and Joe Nudell and Minnie Quartey and Zion Mengesha and Connor Toups and John R. Rickford and Dan Jurafsky and Sharad Goel , title =. Proceedings of the National Academy of Sciences , volume =. 2020 , doi =

2020
[58]

The New York Times , year =

Isaac, Mike , title =. The New York Times , year =
[59]

2023 , publisher=

Selling the American people: Advertising, optimization, and the origins of AdTech , author=. 2023 , publisher=

2023
[60]

Journal of Information Policy , volume=

Notice after notice-and-consent: why privacy disclosures are valuable even if consent frameworks aren't , author=. Journal of Information Policy , volume=. 2019 , publisher=

2019
[61]

2025 , month = oct, day =

Boboris, Kat , title =. 2025 , month = oct, day =

2025
[62]

AI and Ethics , volume=

Let stochastic parrots squawk: Why academic journals should allow large language models to coauthor articles , author=. AI and Ethics , volume=. 2025 , publisher=

2025
[63]

2024 , journal =

Alex Csiszar , title =. 2024 , journal =

2024
[64]

Science , volume =

Keigo Kusumegi and Xinyu Yang and Paul Ginsparg and Mathijs de Vaan and Toby Stuart and Yian Yin , title =. Science , volume =. 2025 , doi =

2025
[65]

2026 , eprint=

Authenticated Contradictions from Desynchronized Provenance and Watermarking , author=. 2026 , eprint=

2026
[66]

Meet Your New Assistant: Meta AI, Built With Llama 3 , year =
[67]

Meet TikTok Symphony, Our New Creative AI Suite , year =
[68]

Nature , year =

Chawla, Dalmeet Singh , title =. Nature , year =. doi:10.1038/d41586-026-01595-5 , url =

doi
[69]

arXiv preprint arXiv:2605.12456 , year=

TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection , author=. arXiv preprint arXiv:2605.12456 , year=

arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Intelligencer (New York Magazine) , year =

Max Read , title =. Intelligencer (New York Magazine) , year =

work page

[2] [2]

Deepfakes and the Epistemic Apocalypse , volume =

Joshua Habgood. Deepfakes and the Epistemic Apocalypse , volume =. doi:10.1007/s11229-023-04097-3 , journal =

work page doi:10.1007/s11229-023-04097-3

[3] [3]

2025 , url =

SynthID: A tool to watermark and identify AI-generated content , howpublished =. 2025 , url =

work page 2025

[4] [4]

2023 , month = nov, day =

Executive Order 14110: Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. 2023 , month = nov, day =

work page 2023

[5] [5]

2025 , month = jan, day =

Executive Order 14179: Removing Barriers to American Leadership in Artificial Intelligence. 2025 , month = jan, day =

work page 2025

[6] [6]

942 (California AI Transparency Act)

Senate Bill No. 942 (California AI Transparency Act). 2024 , month = sep, day =

work page 2024

[7] [7]

An Act regulating provenance regarding artificial intelligence (House Bill No. 90). 2025 , month = feb, day =

work page 2025

[8] [8]

and Miller, Matt L

Cox, Ingemar J. and Miller, Matt L. , urldate =. The First 50 Years of Electronic Watermarking , volume =. 2002 , langid =. doi:10.1155/S1110865702000525 , pages =

work page doi:10.1155/s1110865702000525 2002