Generative AI is destabilizing the foundational assumption of intellectual property law: that creation is difficult, is expensive, and requires legal inducement. When machines produce text, images, and code at near-zero marginal cost, the utilitarian justification for copyright and patent protection begins to collapse. This Essay argues that what emerges in its place is a regime organized not around the scarcity of creation, but around the scarcity of verification—a shift from an incentive paradigm to a source identification paradigm dominated by trademark, rights of publicity, and platform-controlled authentication infrastructure.
OnlyFans provides a revealing case study. Despite AI-generated pornography flooding the market, the platform paid over $5.8 billion to human creators in 2024. Consumers pay not for content, which AI can approximate, but for provenance—the verified knowledge that they are interacting with a specific, authenticated person. This Essay identifies a “triple-lock” structure underlying this economy: verified identity draws consumers in, proprietary infrastructure ensures that access requires staying inside the walled garden, and aggressive copyright enforcement destroys unauthorized copies that would otherwise undermine authenticity’s value.
Applying a Law and Political Economy lens, the Essay argues that this emerging regime systematically advantages those who already possess recognized brands, legal departments, and capital for proprietary infrastructure—while offering little to individual creators who lack pre-existing fame. It concludes by proposing that the Library of Congress serve as a public “digital notary,” preventing the infrastructure of authenticity from becoming a private toll road.
Introduction
AI-generated pornography is flooding the internet; The Economist reports that the market for synthetic adult content will reach $2.5 billion in 2025 and is projected to grow at twenty-seven percent annually.1 1.AI Is Upending the Porn Industry, The Economist, Nov. 29, 2025, at 55, 55.Show More And yet, OnlyFans, the online adult content platform designed around human content creators, generated $7.22 billion in revenue in 2024.2 2.Id. at 56; Todd Spangler, OnlyFans Gross Revenue Rises 9% to $7.2 Billion in 2024, Variety (Aug. 22, 2025, at 05:01 PT), https://variety.com/2025/digital/news/onlyfans-fiscal-2024-revenue-earnings-1236495750/.Show More The platform has paid over $25 billion to creators since 2016.3 3.Rose Henderson, OnlyFans Has Paid Creators $25 Billion Since 2016, CEO Says, Bloomberg (Oct. 21, 2025, at 11:06 ET), https://www.bloomberg.com/news/articles/2025-10-21/onlyfans-has-paid-creators-25-billion-since-2016-ceo-says.Show More
These facts sit uneasily together. If generative AI can produce unlimited synthetic explicit content on demand, customized to any specification, and cheaply available—why would anyone pay for content from a specific human being? The answer illuminates something important about the future of intellectual property. What consumers pay for on OnlyFans is not the content itself, which AI can now approximate with increasing fidelity. They pay for the provenance—the verified knowledge that they are interacting with a specific, authenticated person. The content is abundant; the source is scarce.
But provenance alone is not enough. OnlyFans does not merely verify identity; it protects that identity within a fortress of interlocking legal and technological defenses. Verified identity draws consumers in. Proprietary infrastructure—DRM encryption, anti-scraping measures, behavioral data collection—functions as enclosure. And aggressive copyright enforcement destroys unauthorized copies that would otherwise flood the market with “authentic” content. This triple-lock structure—identity as the hook, infrastructure as the lock, enforcement as border control—is the architecture of the post-incentive economy.
The intellectual property system has long operated on a different premise: that creation is difficult, expensive, and scarce, and therefore creators require the inducement of a temporary monopoly to produce. This utilitarian bargain underwrites the patent and copyright clause of the U.S. Constitution.4 4.U.S. Const. art. I, § 8, cl. 8.Show More Yet generative artificial intelligence has begun to unravel this foundational assumption. When the marginal cost of producing competent text, images, and code approaches zero, the economic logic of incentivizing creation collapses. Something must fill the void.
What emerges is a regime organized not around the scarcity of creation, but around the scarcity of verification. We are witnessing a transition from what might be called the incentive paradigm of intellectual property to a source identification paradigm dominated by trademark, rights of publicity, and platform-controlled verification infrastructure. In this new order, the legal system’s primary function shifts from encouraging the production of goods to certifying their origin.
This transformation is often presented as a neutral response to technological change.5 5.See, e.g., C2PA Founding Press Release, Coal. for Content Provenance & Authenticity (Feb. 22, 2021), https://c2pa.org/c2pa-founding-press-release/ [https://perma.cc/47CF-75QN] (“With the collective expertise of this group, we will accelerate the critical work of rebuilding the public’s trust in online content through broad and open adoption of a provenance standard at scale.”); Exec. Order No. 14110, 3 C.F.R. 657 (2024) (treating content provenance and watermarking standards as self-evidently beneficial transparency measures, without addressing governance of the underlying infrastructure); Council Regulation 2024/1689, ch. IV, art. 50, ¶ 2, 2024 O.J. (L) 82 (EU) (mandating machine-readable labeling of AI-generated outputs that are “effective, interoperable, robust and reliable” without note of the infrastructure ownership).Show More This Essay argues that such framing obscures the distributive stakes. The emerging regime favors those who already possess recognized brands, access to legal departments, and the capital to build proprietary infrastructure. It promises to entrench the dominance of large platforms and legacy media organizations while offering little to individual creators who lack pre-existing fame.6 6.For the Law and Political Economy framework applied here, see Jedediah Britton-Purdy, David Singh Grewal, Amy Kapczynski & K. Sabeel Rahman, Building a Law-and-Political-Economy Framework: Beyond the Twentieth-Century Synthesis, 129 Yale L.J. 1784 (2020).Show More
While legal scholarship has extensively diagnosed the erosion of the incentive paradigm, it has yet to fully articulate replacing it. Mark Lemley has persuasively argued that the digital age renders IP’s artificial scarcity inefficient,7 7.Mark A. Lemley, IP in a World Without Scarcity, 90 N.Y.U. L. Rev. 460, 470–71 (2015).Show More and Pamela Samuelson has warned against rushing to create sui generis rights for AI outputs.8 8.See Pamela Samuelson, Generative AI Meets Copyright, 381 Sci. 158, 160–61 (2023).Show More Yet, the critical question remains unanswered: if the economic logic of incentivizing creation is collapsing, how is value being enclosed in its absence? This Essay provides the first unified account of that replacement regime, offering both a diagnostic tool for the post-incentive economy and a prescriptive path to prevent the authentication regime from becoming a form of private enclosure.
This Essay proceeds in four Parts. Part I documents the collapse of the incentive paradigm. Part II analyzes the emerging regime of provenance and private control. Part III turns to political economy, arguing that the triple-lock structure systematically advantages incumbents over entrants. Part IV proposes that the Library of Congress could serve as a public root of trust—a “digital notary” that prevents the infrastructure of authenticity from becoming a private toll road.
- AI Is Upending the Porn Industry, The Economist, Nov. 29, 2025, at 55, 55. ↑
- Id. at 56; Todd Spangler, OnlyFans Gross Revenue Rises 9% to $7.2 Billion in 2024, Variety (Aug. 22, 2025, at 05:01 PT), https://variety.com/2025/digital/news/onlyfans-fiscal-2024-revenue-earnings-1236495750/. ↑
- Rose Henderson, OnlyFans Has Paid Creators $25 Billion Since 2016, CEO Says, Bloomberg (Oct. 21, 2025, at 11:06 ET), https://www.bloomberg.com/news/articles/2025-10-21/onlyfans-has-paid-creators-25-billion-since-2016-ceo-says. ↑
- U.S. Const. art. I, § 8, cl. 8. ↑
- See, e.g., C2PA Founding Press Release, Coal. for Content Provenance & Authenticity (Feb. 22, 2021), https://c2pa.org/c2pa-founding-press-release/ [https://perma.cc/47CF-75QN] (“With the collective expertise of this group, we will accelerate the critical work of rebuilding the public’s trust in online content through broad and open adoption of a provenance standard at scale.”); Exec. Order No. 14110, 3 C.F.R. 657 (2024) (treating content provenance and watermarking standards as self-evidently beneficial transparency measures, without addressing governance of the underlying infrastructure); Council Regulation 2024/1689, ch. IV, art. 50, ¶ 2, 2024 O.J. (L) 82 (EU) (mandating machine-readable labeling of AI-generated outputs that are “effective, interoperable, robust and reliable” without note of the infrastructure ownership). ↑
- For the Law and Political Economy framework applied here, see Jedediah Britton-Purdy, David Singh Grewal, Amy Kapczynski & K. Sabeel Rahman, Building a Law-and-Political-Economy Framework: Beyond the Twentieth-Century Synthesis, 129 Yale
L.J.
1784 (2020). ↑
- Mark A. Lemley, IP in a World Without Scarcity, 90 N.Y.U. L. Rev
.
460, 470–71 (2015). ↑
- See Pamela Samuelson, Generative AI Meets Copyright, 381 Sci. 158, 160–61 (2023). ↑
Click on a link below to access the full text of this article. These are third-party content providers and may require a separate subscription for access.