Majority of ‘In Re: OpenAI, Inc. Copyright Infringement Litigation’ Amicus Briefs Support Plaintiff Copyright Owners
As the Southern District of New York considers summary judgment in In Re: OpenAI, Inc. Copyright Infringement Litigation, 52 amicus briefs were filed last week by industry groups, academics, artists, creative unions, startup AI companies, government officials and many others. Nearly 30 of those briefs support the plaintiffs (or oppose the defendant AI companies), urging the court to reject various arguments by the defendants, their supporters, and the Department of Justice in its Statement of Interest. Below are a few key takeaways:
- Amicus briefs filed in support of plaintiffs explain that AI model training is non-transformative because models consume copyright-protected expressive works to generate commercial market substitutes.
- A diverse range of amici detail how uncompensated AI scraping directly harms established and emerging licensing markets that already demonstrate commercial viability.
- Many of the briefs warn that permissionless ingestion strips critical advertising and subscription revenues from publishers, creators, and local media.
- Small and startup AI companies explain that licensing is not only possible but essential to build safe and reliable models that benefit all stakeholders.
- Members of Congress and advocacy organizations warn that the Department of Justice’s Statement of Interest does not represent the position of Congress and in fact is at odds with the Administration’s stated positions on AI and IP.
Notable Amicus Briefs
In Support of Plaintiffs
AI Coalition for Data Integrity, joined by the Writers Guild of America West and East, National Association of Voice Actors, Microstocksolutions, Digital Media Licensing Association, Transparency Coalition AI, Association of Medical Illustrators, Association of Independent Music Publishers, and Music Workers Alliance, submitted a brief arguing that national security and foreign legal standards disfavor undocumented AI training. They emphasize that a legitimate training-data licensing market already exists and urge the court to analyze each distinct operational use separately based on verifiable records of what was copied.
AI Rights Project submitted a brief arguing that claims that AI training is transformative fair use are based on the flawed premise that machine-scale AI training is comparable to human learning, even though AI systems extract protected expressive details far beyond human cognitive limits. The brief further contends that this human baseline is crucial to market-harm analysis, and that the court should not disregard statutory rights in favor of general public benefits. It also argues that Congress, not the judiciary, should make any significant change to copyright law to address machine-scale training.
AI Startups including BandLab, Boomy Corporation, Created by Humans, Credtent, Distributed Creation, Newstex, ProrataAI, SourceAudio Holdings, Vandal AI, and Voice-Swap, submitted a brief arguing that developers can build commercially useful AI products without violating copyright law because piracy-free training is already feasible, commercially viable, and will not price out smaller developers. They further argue that licensing supports innovation and distributes AI’s economic benefits by creating new market opportunities, rewarding original human creation, reducing litigation risk, and ensuring prosperity is shared beyond large frontier-model developers.
America’s Newspapers submitted a brief arguing that AI developers have built models using copied local reporting specifically to replace original outlets, resulting in particularly severe losses of traffic and revenue for local publishers. It further argues that the decline of local news increases government costs, reduces accountability, diminishes public safety, and undermines American democracy—harms that far outweigh any purported public benefits of AI.
American Federation of Musicians, joined by the Screen Actors Guild–American Federation of Television and Radio Artists (SAG-AFTRA) and the Writers Guild of America West and East, submitted a brief arguing that the news and entertainment industries rely on authorized licensing to sustain creative livelihoods. The brief asserts that fair use analyses cannot separate AI model training from its purpose of memorizing human expression and its impact on competitive markets. It concludes that robust licensing markets better serve the public interest than a categorical fair use exemption.
Ashley Judd submitted a brief arguing that copyright incentives are vital for financing investigative journalism, which relies on personal trust, editorial judgment, and human relationships between reporters and sources. The brief explains that generative AI products threaten this ecosystem by producing unauthorized narrative substitutes, diverting news revenues, risking hallucinated source accounts, and amplifying historical biases embedded in training data.
Association of American Publishers, joined by Copyright Clearance Center, Independent Book Publishers Association, and International Association of Scientific, Technical & Medical Publishers, submitted a brief arguing that OpenAI’s copying for AI training is not transformative, as language models utilize expressive text for the same purposes as the original works and produce market substitutes. Additionally, amici argue that unlicensed AI training causes substantial market harm by undermining both original and derivative works, displacing the emerging AI licensing market, and fostering piracy.
Centre for Technology & Democracy, joined by affiliated scholars, Director Gina Neff, and Taryn J.M. Ward, submitted a brief arguing that key aspects of the nature of defendants’ use of copyrighted works represent disputed, material facts. It further argues that AI training is an ongoing process in which copyrighted expression remains embedded in model parameters. Importantly, the brief argues that because expressive encodings persist in finished systems, AI training cannot be considered temporary copying or protected by intermediate copying defenses.
Copyright Alliance submitted a brief explaining that “intermediate copying” for AI model development does not serve an ultimately transformative purpose. Still, even if it did, the brief explains that a finding of transformativeness does not control the fair use analysis. The brief also demonstrates that a robust licensing market for AI training exists, and that emerging licensing markets can be harmed even when a copyright owner has not yet entered the market. Addressing the past fair use cases the defendants rely on, the brief explains that they are clearly distinguishable from the OpenAI dispute, especially in light of the U.S. Supreme Court’s recent decision in Warhol v. Goldsmith.
Digital Content Next submitted a brief arguing that digital web traffic is essential to publisher survival, that the unauthorized use of copyrighted works by AI developers destroys critical advertising and subscription revenue, that market substitution defeats defendants’ fair-use defense, and that publisher silence does not create an implied license to scrape content. The brief warns that permissionless scraping creates a destructive “doom loop” for the information economy.
First Amendment Scholars, including RonNell Andersen Jones, Jack M. Balkin, Erin Carroll, Alan Chen, Heidi Kitrosser, Jane E. Kirtley, Burt Neuborne, Jonathan Peters, Victor Pickard, Francesca Procaccini, and Sonja R. West, submitted a brief arguing that the independent press is essential to democracy by holding power to account, and that defendants’ unauthorized use of copyrighted reporting threatens journalism’s economic survival. The brief maintains that AI tools cannot replace a free press and warns that allowing uncompensated use of the fruits of news labor undermines constitutional democracy.
Jim Albrecht, former Senior Director of News Ecosystem Products at Google, submitted a brief arguing that Retrieval-Augmented Generation (RAG) engages in industrial-scale copying beyond the scope of implied web licenses, systematically extracting news content and creating substitute products that divert web traffic and revenue from publishers. The brief further asserts that AI developers cannot raise a legitimate fair use defense by framing model transformations as human-like rewriting while ignoring the underlying unauthorized verbatim copying.
Joseph Stiglitz, a Nobel Prize-winning economist and Columbia University professor, submitted a brief arguing that generative AI intermediaries disrupt the economic ecosystem for reliable information by diverting reader attention and revenue, and that fair use should be limited because commercial AI products directly substitute for original works. The brief warns that unauthorized AI ingestion lowers the cost of misinformation and deprives original producers of resources, ultimately degrading society’s knowledge base and risking an information collapse.
Ken Burns, documentary filmmaker, submitted a brief arguing that the defendants’ unauthorized use of copyrighted documentary scripts to train generative AI models undermines the market for quality historical documentaries by enabling the creation of cheap, inferior copies. The brief further contends that AI models produce verbatim, non-transformative replicas of protected works, which do not qualify as fair use.
News/Media Alliance, joined by the National Newspaper Association, the National Press Photographers Association, and ProPublica, submitted a brief arguing that news organizations depend on copyright protection to fund journalism, and that unauthorized copying for generative AI model development creates non-transformative commercial substitutes that severely harm both established online news markets and emerging AI licensing ecosystems. The brief warns that letting technology companies exploit journalistic works without compensation threatens the economic viability of an independent press and risks an informational “doom loop,” underscoring the need for strong copyright enforcement and market-based licensing to protect the public interest.
Open Markets Institute submitted a brief arguing that the defendants’ large-scale, unauthorized copying of news content constitutes unfair and socially destructive competition by creating market substitutes that strip publishers of essential subscription and advertising revenue. It further argues that this uncompensated ingestion penalizes law-abiding AI rivals who pay for content, generating an illegitimate cost advantage that risks destroying the emerging licensing market for copyrighted material.
PEN America, joined by the Protect Democracy Project and Reporters Without Borders, submitted a brief arguing that the Constitution’s Press Clause and Progress Clause were intended to work together to support democratic self-government by encouraging valuable writing and informing the public. The brief contends that the defendants’ fair use claims undermine this constitutional balance by threatening the independent press’s economic survival and replacing resource-intensive journalism with commercial AI substitutes.
Photographers, Playwrights, and Literary Translators, represented by American Society of Journalists and Authors, American Society of Media Photographers, American Literary Translators Association, Garden Communications International, National Association of Science Writers, National Press Photographers Association, National Writers Union, Novelists, Romance Writers of America, Sisters in Crime, The Dramatists Guild of America, and the Science Fiction and Fantasy Writers Association, submitted a brief arguing that the fourth fair use factor is paramount, and that fair use should not apply when unauthorized uses are highly commercial, minimally transformative, conducted in bad faith, or intended to replace human creative works. The brief warns that by flooding markets with AI-generated substitutes, diverting revenue from human creators, and exploiting unauthorized copyrighted material, the defendants significantly undermine economic incentives to create and fail to meet fair use requirements.
Progressive Democrats of America, joined by AI Safety Asia, Brian J. A. Boyd, Buddhism & AI Initiative, Cara Project, Charles Camosy, David Atkinson, EdTech Law Center PLLC, Future of Life Institute, Irreplaceable, John C. Havens, Noēsis Collaborative, RadicalxChange Foundation, The Tech Oversight Project, Walter Scheirer, and Young People’s Alliance, submitted a brief arguing that unauthorized generative AI training does not serve copyright’s core purpose and that treating model learning as fair use would improperly grant AI privileges not afforded to human learners. The brief further contends that publicly available content does not imply a license for unauthorized scraping, and that granting special status to generative AI threatens the web’s economy and broader human progress.
Rebuild Local News, joined by Local Independent Online News Publishers, The NewsGuild-CWA, the National Association of Hispanic Publications, Solutions Journalism Network, AAN Publishers, the Institute for Rural Journalism and Community Issues, and The Society of Professional Journalists, submitted a brief arguing that uncompensated AI scraping drains essential web traffic, advertising, and subscription revenue from local media, leading to the decline of local news and causing significant civic, economic, and democratic harm. The brief explains that rejecting fair use defenses and enforcing strong copyright protections will force AI companies to adopt fair licensing practices, helping preserve community journalism and prevent a downward spiral of misinformation.
Recording Industry Association of America, joined by National Music Publishers Association, American Association of Independent Music, Songwriters of North America, Black Music Action Coalition, National Academy of Recording Arts & Sciences, American Society of Composers, Authors and Publishers, Broadcast Music, and Artist Rights Alliance, submitted a brief arguing that training AI models on copyrighted works to create synthetic substitutes undermines copyright’s core purpose of encouraging human authorship, fails the first fair use factor due to its non-transformative nature, and causes significant market harm by producing commercial substitutes. The brief further argues that because machine-generated outputs lack human authorship and displace human creators in both new and traditional markets, unauthorized training cannot qualify as fair use.
William Morris Endeavor submitted a brief arguing that a robust market already exists for licensing copyrighted works and creator personas for AI training, and that unauthorized AI ingestion causes direct market harm by creating substitute works, replacing performer likenesses, and depriving rightsholders of licensing fees. The brief maintains that requiring developers to pay for creative material will not hinder innovation, just as licensing adapted to radio and streaming, and that uncompensated AI ingestion severely damages established licensing markets under the fourth fair use factor.
In Support of Neither Party
Center for American Rights, joined by Article III Project, Internet Accountability Project, The Bull Moose Project, Consumer Action for a Strong Economy (CASE), and Joel Thayer, submitted a brief arguing that adopting the defendant’s broad fair-use defense would upset the balance between AI innovation and copyright protection, since uncompensated data scraping undermines creator incentives and endangers media diversity. The brief also contends that robust intellectual property enforcement is vital for national security, warning that an expansive fair-use standard could enable foreign adversaries to legally steal IP and would threaten U.S. technological leadership.
Center for AI and Digital Policy, joined by AI experts and legal scholars, submitted a brief arguing that effective AI governance depends on training-data transparency and that copyright law should evaluate specific operational acts, not encourage opaque development or poor recordkeeping. The brief also maintains that thorough dataset documentation both protects rights and promotes responsible innovation.
Copyright Law Professors, including Robert Brauneis, Sandra Aistars, Mary Catherine Amerine, Samuel F. Ernst, Vicenç Feliú, Jon Garon, Thomas Haley, Justin Hughes, Deirdré A. Keller, Yolanda M. King, Michael Landau, Yvette Joy Liebesman, Adam Mossoff, Loren E. Mulraine, Christopher Newman, Seán O’Connor, Dr. David W. Opderbeck, Dr. Julia Powles, Trevor G. Reed, Marketa Trimble, and Bhamati Viswanathan, submitted a brief arguing that defendants’ unauthorized copying of copyrighted expressive works to train computer systems that outperform human authors constitutes direct copyright infringement, since protected expression is inseparable from underlying facts and ideas in the plaintiffs’ works. The brief further argues that such ingestion is not protected by fair use because its commercial and minimally transformative purpose causes significant market harm and does not qualify as permissible intermediate copying or non-expressive use.
High School Students in Los Angeles County submitted a brief offering a perspective from the next generation of copyright holders who are “coming of age of age in a digital world that increasingly treats our creative labor not as property to be protected, but as raw material to be mined.” Calling it “extraction,” the brief argues that AI training is not transformative because it simply ingests works for the purpose of mimicking them. The brief goes on to warn of the dangers an AI “doom loop” that would threaten the future of legitimate human journalism, and it concludes by drawing attention to the fact that AI companies’ unchecked copying is harming education and diminishing incentives for human creativity.
Human Consent Foundation, joined by the Creative Artists Agency, Music Artists Coalition, National Association of Voice Actors, Writer, Creators Coalition on Artificial Intelligence, Cate Blanchett, Nick Hexum, and Steven Soderbergh, submitted a brief arguing that collective licensing of dispersed rights holders is well established, that machine-readable consent infrastructure is already feasible and in use, and that an operational AI licensing market exists for premium content and is expanding into broader categories. The brief further argues that the fourth fair use factor should reflect these market realities rather than rely on assumptions of transactional impossibility, to protect rightsholders’ fundamental statutory right to affirmative opt-in consent.
Members of Congress, including Representatives Jamie Raskin (D-MD), Jerrold Nadler (D-NY), and Henry C. “Hank” Johnson, Jr. (D-GA), submitted a brief arguing that the Department of Justice improperly invokes national security in copyright law, claiming it is not a statutory fair use factor under Section 107 and that the agency lacks authority to create such a broad exemption. The brief further asserts that this approach reflects unilateral agency action, bypasses the U.S. Copyright Office’s expertise, and overlooks copyright’s essential economic and cultural role, concluding that the court should give little weight to the Statement of Interest.
Michael Moritz, a technology investor, philanthropist, author, and former journalist, submitted a brief explaining that there is a workable licensing market between publishers and the developers of large language models and urging the court find that LLM developers cannot copy the works of journalists and offer substitutive news without offering any form of compensation. Focusing on the future of independent journalism, the brief stresses that licensing markets exist for AI training and are essential for both journalism and AI industries to thrive.
Conclusion
The diversity of individuals and organizations filing briefs in support of the news publisher and book author plaintiffs is striking. Many amici agree that unauthorized AI model ingestion amounts to non-transformative commercial copying that causes severe market harm by creating substitutes, diverting revenue from creators and publishers, and weakening viable licensing markets. Although the defendants in the case were supported by (fewer) amici that argue that uncompensated scraping is essential to technological innovation and national security, that argument rests on the flawed assumption that machine ingestion equals human learning and disregards workable licensing models. Hopefully, the Southern District of New York will recognize the real risks to the creative community and issue a fair use decision that thoughtfully balances innovation with the rights of creators. This blog highlights only a selection of the many briefs submitted in this case, with additional briefs available here.
To stay up to date with the latest news in artificial intelligence (AI) and copyright, sign up for our AI Copyright Alert. You can also visit our AI and Copyright hub for additional resources on federal court cases, current licensing, and more. Additionally, if you aren’t already a member of the Copyright Alliance, you can join today by completing our Creator Membership form! Members gain access to monthly newsletters, educational webinars, and so much more — all for free!