Phelps Dunbar LLP Logo
  • Services
  • Insights
  • Professionals
Phelps Dunbar LLP Logo
  • Services
  • Insights
  • Professionals
  • ABOUT US
  • LOCATIONS
  • SUSTAINABILITY
  • CAREERS
  • Practices
  • Industries

    Recent Firings and Resignations Show the Depth of Disputes Over Copyright Fair Use and Training Gen AI

    May 22, 2025

    This article was written and published for Law.com.

    The largest copyright dispute in history may be whether training generative artificial intelligence (Gen AI) models is copyright fair use. It is a question that divides academics, practitioners, and industries. For content creation companies like movie studios, publishers, record labels, and the news media, the use of their copyrighted works to train Gen AI models looks like theft. For tech companies, the use of copyrighted material to train Gen AI models is no different from authors reading others’ books to improve their own craft.

    In May 2025, that dispute came to a head, apparently resulting in the termination of the Librarian of Congress and the Registrar of Copyright, and the resignation of prominent copyright law professors from the American Law Institute’s (ALI) Copyright Restatement Project.

    At the Copyright Office, those firings may have been related to the administration’s support of a broad fair use policy that would allow the rapid expansion of Gen AI technology. In academia, the resignations were driven by just the opposite concern. There, the individuals who resigned were concerned that the proposed Restatement too strongly supports the finding of fair use for the use of copyright materials in training Gen AI models.

    What does this mean? Ultimately, the courts (and presumably the U.S. Supreme Court) will decide the issue of fair use. However, others have the ability to create or influence policy. The firings, for example, may mean that the administration is at least considering finding that using copyrighted materials in training Gen AI models is necessary for national security.

    The Firings

    On Thursday May 8, the Librarian of Congress, who oversees the Copyright Office, was fired by President Donald Trump. The next day, Friday May 9, the Copyright Office took the unusual step of making a “pre-publication” release of the third portion of its report on copyright and AI. That report took a middle ground position on Gen AI and fair use, finding that each decision regarding whether using copyrighted material to train Gen AI models was a fair use would be dependent on the underlying facts. The released report is not official Copyright Office policy yet and may have been released in its pre-publication form to ensure the report made its way to the public before political pressure required a change in the report’s findings.

    Following the release of the pre-publication report the previous day, on Saturday, May 10, it was reported that Trump had fired the Register of Copyrights, the person who runs the day-to-day operations of the Copyright Office. On Monday May 12, the Department of Justice confirmed that Todd Blanche had been appointed acting Librarian of Congress. The DOJ declined to comment regarding the appointment of an Acting Register of Copyrights.

    One reason for that declination is that it is not clear the president can fire the Register of Copyrights. As the name suggests, the Library of Congress is part of the legislative branch, not the executive branch. While the Librarian of Congress is nominated by the President and confirmed by the Senate, it is the Librarian who has the power to hire and fire the Register of Copyrights.

    Copyright Lately has an excellent analysis of why the existence of the draft report may have been responsible for the firings at the Copyright Office. In wrapping up his breakdown of these events, Aaron Moss writes:

    “The Copyright Office isn’t picking winners or losers in the 40-plus AI copyright cases now pending in court, and its report repeatedly emphasizes that fair use turns on the specific facts of each case. But taken as a whole, the analysis reads as broadly favorable to copyright owners—most notably in its endorsement of a novel market dilution theory that no court has yet adopted.”

    Other outlets have reported Elon Musk may have the president’s ear on this issue. That is important as one of X’s latest innovations is the broad deployment of Grok, an Gen AI chatbox. For tech companies like X, the difference between a finding of fair use and a finding of infringement could mean billions of dollars in losses and delays in moving forward with new Gen AI technology.

    The Resignations

    The Monday following the firings, resignations from ALI began. ALI gathers eminent law professors, practitioners, and thought leaders to draft, revise, and publish Restatements of the Law, that are enormously influential in the courts and legislatures, as well as in legal scholarship and education. Restatements are supposed to be statements of what the law is in a particular area, along with supporting authorities, and expert analysis. Currently, ALI is working on a revised Restatement of Copyright Law.

    On Monday May 12, Profs. Shyam Balganesh, Jane Ginsburg and Peter Menell resigned from the project, writing: “In the interests of copyright law, processual integrity, and the continuing success of the Institute, we hereby resign from the project, and request that our names be removed from the list of advisers.” Joining them in resigning was David Nimmer, the author of the most influential treatise on copyright law. On Friday May 16, Copyright Alliance CEO Keith Kupferschmid also resigned. Kupferschmid stated,

    “I am officially announcing my departure from the ALI’s Copyright Restatement project for a number of critical reasons, including the fact that throughout the initiative, ALI Copyright Restatement Reporters routinely disregarded concerns and comments shared by the U.S. Copyright Office, judges, and numerous other expert participants. This input was not considered because it differed from the ALI reporters’ views and biases regarding copyright law. Before the start of this project, we and many others warned that the Copyright Restatement effort was likely to look much more like a restatement of the reporters’ views rather than actual copyright law. Unfortunately, these predictions came true, resulting in the final version of the Copyright Restatement presenting a restrictive and warped view of copyright—one that is inconsistent with the statute and case law and reflects the reporters’ aspirations to change the copyright law. There is a general undercurrent of anti-copyright sentiment that runs through the entire Copyright Restatement that manifests itself through a disproportionate focus on atypical court decisions that limit the scope of copyright protection.”

    Based on their public statements, it appears that the concerns of those who resigned were related, at least in part, to a dispute over how the Restatement should treat the issue of fair use, especially as it concerns Gen AI training models. In fact, the group led by Kupferschmid has filed an amicus brief arguing that Meta’s use of copyrighted material to train its Gen AI models is not fair use.

    By May 16, the dispute among the ALI advisers had reached social media. Prof. Pamela Samuelson, took to X to defend the work on the Restatement in a series of posts. Samuelson went as far as calling those who resigned “sore losers.” Re:Create Executive Director Brandon Butler, who also served as an adviser to the project, wrote that he had “nothing but admiration for the [Restatement’s] author’s extraordinary patience, humility, and erudition in the face of a ceaseless onslaught of sniping and mean-spirited, unfair attacks from critics, including some advisers.”

    There has not yet been significant reporting on the ALI deliberations, but there is a clear split in academia over the substantive issue. Ten law professors filed an amicus brief on behalf of authors, arguing that Meta’s use of copyrighted material to train Gen AI models was not fair use. In the same case, four law professors filed an amicus brief on behalf of Meta’s argument.

    On May 20, ALI’’s membership adopted the tentative Restatement by a voice vote.

    Why it Matters?

    With billions on the line and certain industries, such as the news media, imperiled by the rise of Gen AI, the stakes of the fair use question could not be higher. That makes it obvious industry, on both sides of the issue, would work to ensure the most support for their position. Two places to assert such influence are at the Copyright Office and ALI.

    Both the Copyright Office and ALI provide important guidance for courts and practitioners. However, courts are not required to follow either the Copyright Office or the Restatement. Instead, it is the quality of the work these experts produce that makes their work products persuasive. Thus, these firings and resignations may do nothing more than serve as a warning to courts to carefully consider the reasoning in the Copyright Office’s report and the proposed revision to the Restatement.

    More importantly, the public playing-out of this dispute may signal what actions the administration will take regarding fair use. While only Congress can update copyright law, the executive is not powerless.

    The White House is currently forming an AI Action Plan that “will define priority policy actions to enhance America’s position as an AI powerhouse and prevent unnecessarily burdensome requirements from hindering private sector innovation.” As part of the public comments to that plan, OpenAI proposed that the President should declare that using copyrighted materials to train Gen AI is allowable because developing Gen AI technology is important to national security. While such a position is obviously beneficial to OpenAI’s business, it is not without support. As OpenAI points out, China has “committed to ambitious targets for AI adoption across its public administration, security, and the military.”

    This national security angle could encourage the President to override the Copyright Act to allow for the use of copyrighted materials in training Gen AI models. If he made such a decision, Trump would have not only the support of much of the tech industry, but also from at least some in academia.

    Prof. Matthew Sag, a law professor specializing in artificial intelligence, machine learning, and data science at Emory University, also submitted a comment to the AI Action Plan. Sag has explicitly stated that the president does not have the power to override copyright law by executive order, but has also expressly stated that national security is at risk if Gen AI training models cannot use copyright materials. As Sag put it, a “bad court decision may drive AI innovation offshore.”

    What's Next

    Disputes at the Copyright Office and ALI show that there is no broad consensus regarding the right answer to the fair use and training model question. Courts are an excellent vehicle for resolving disputes between parties and interpreting laws. Courts are not built to craft broad policies regarding how GenAI will be built in the United States without harming authors, filmmakers, musicians, and other content creators. Ultimately, that is a job for legislators. If Congress does not act, then the administration may well try to mold copyright policy to its own goals. No matter the position Congress or the President take, they will find ample support and opposition in academia and industry. For now, watch closely to see who becomes the Register of Copyrights and how the ALI reacts to the recent resignations.

     

    One reason for that declination is that it is not clear the president can fire the Register of Copyrights. As the name suggests, the Library of Congress is part of the legislative branch, not the executive branch. While the Librarian of Congress is nominated by the President and confirmed by the Senate, it is the Librarian who has the power to hire and fire the Register of Copyrights.

    Copyright Lately has an excellent analysis of why the existence of the draft report may have been responsible for the firings at the Copyright Office. In wrapping up his breakdown of these events, Aaron Moss writes:

    “The Copyright Office isn’t picking winners or losers in the 40-plus AI copyright cases now pending in court, and its report repeatedly emphasizes that fair use turns on the specific facts of each case. But taken as a whole, the analysis reads as broadly favorable to copyright owners—most notably in its endorsement of a novel market dilution theory that no court has yet adopted.”

    Other outlets have reported Elon Musk may have the president’s ear on this issue. That is important as one of X’s latest innovations is the broad deployment of Grok, an Gen AI chatbox. For tech companies like X, the difference between a finding of fair use and a finding of infringement could mean billions of dollars in losses and delays in moving forward with new Gen AI technology.

    The Resignations

    The Monday following the firings, resignations from ALI began. ALI gathers eminent law professors, practitioners, and thought leaders to draft, revise, and publish Restatements of the Law, that are enormously influential in the courts and legislatures, as well as in legal scholarship and education. Restatements are supposed to be statements of what the law is in a particular area, along with supporting authorities, and expert analysis. Currently, ALI is working on a revised Restatement of Copyright Law.

    On Monday May 12, Profs. Shyam Balganesh, Jane Ginsburg and Peter Menell resigned from the project, writing: “In the interests of copyright law, processual integrity, and the continuing success of the Institute, we hereby resign from the project, and request that our names be removed from the list of advisers.” Joining them in resigning was David Nimmer, the author of the most influential treatise on copyright law. On Friday May 16, Copyright Alliance CEO Keith Kupferschmid also resigned. Kupferschmid stated,

    “I am officially announcing my departure from the ALI’s Copyright Restatement project for a number of critical reasons, including the fact that throughout the initiative, ALI Copyright Restatement Reporters routinely disregarded concerns and comments shared by the U.S. Copyright Office, judges, and numerous other expert participants. This input was not considered because it differed from the ALI reporters’ views and biases regarding copyright law. Before the start of this project, we and many others warned that the Copyright Restatement effort was likely to look much more like a restatement of the reporters’ views rather than actual copyright law. Unfortunately, these predictions came true, resulting in the final version of the Copyright Restatement presenting a restrictive and warped view of copyright—one that is inconsistent with the statute and case law and reflects the reporters’ aspirations to change the copyright law. There is a general undercurrent of anti-copyright sentiment that runs through the entire Copyright Restatement that manifests itself through a disproportionate focus on atypical court decisions that limit the scope of copyright protection.”

    Based on their public statements, it appears that the concerns of those who resigned were related, at least in part, to a dispute over how the Restatement should treat the issue of fair use, especially as it concerns Gen AI training models. In fact, the group led by Kupferschmid has filed an amicus brief arguing that Meta’s use of copyrighted material to train its Gen AI models is not fair use.

    By May 16, the dispute among the ALI advisers had reached social media. Prof. Pamela Samuelson, took to X to defend the work on the Restatement in a series of posts. Samuelson went as far as calling those who resigned “sore losers.” Re:Create Executive Director Brandon Butler, who also served as an adviser to the project, wrote that he had “nothing but admiration for the [Restatement’s] author’s extraordinary patience, humility, and erudition in the face of a ceaseless onslaught of sniping and mean-spirited, unfair attacks from critics, including some advisers.”

    There has not yet been significant reporting on the ALI deliberations, but there is a clear split in academia over the substantive issue. Ten law professors filed an amicus brief on behalf of authors, arguing that Meta’s use of copyrighted material to train Gen AI models was not fair use. In the same case, four law professors filed an amicus brief on behalf of Meta’s argument.

    On May 20, ALI’’s membership adopted the tentative Restatement by a voice vote.

    Why it Matters?

    With billions on the line and certain industries, such as the news media, imperiled by the rise of Gen AI, the stakes of the fair use question could not be higher. That makes it obvious industry, on both sides of the issue, would work to ensure the most support for their position. Two places to assert such influence are at the Copyright Office and ALI.

    Both the Copyright Office and ALI provide important guidance for courts and practitioners. However, courts are not required to follow either the Copyright Office or the Restatement. Instead, it is the quality of the work these experts produce that makes their work products persuasive. Thus, these firings and resignations may do nothing more than serve as a warning to courts to carefully consider the reasoning in the Copyright Office’s report and the proposed revision to the Restatement.

    More importantly, the public playing-out of this dispute may signal what actions the administration will take regarding fair use. While only Congress can update copyright law, the executive is not powerless.

    The White House is currently forming an AI Action Plan that “will define priority policy actions to enhance America’s position as an AI powerhouse and prevent unnecessarily burdensome requirements from hindering private sector innovation.” As part of the public comments to that plan, OpenAI proposed that the President should declare that using copyrighted materials to train Gen AI is allowable because developing Gen AI technology is important to national security. While such a position is obviously beneficial to OpenAI’s business, it is not without support. As OpenAI points out, China has “committed to ambitious targets for AI adoption across its public administration, security, and the military.”

    This national security angle could encourage the President to override the Copyright Act to allow for the use of copyrighted materials in training Gen AI models. If he made such a decision, Trump would have not only the support of much of the tech industry, but also from at least some in academia.

    Prof. Matthew Sag, a law professor specializing in artificial intelligence, machine learning, and data science at Emory University, also submitted a comment to the AI Action Plan. Sag has explicitly stated that the president does not have the power to override copyright law by executive order, but has also expressly stated that national security is at risk if Gen AI training models cannot use copyright materials. As Sag put it, a “bad court decision may drive AI innovation offshore.”

    What's Next

    Disputes at the Copyright Office and ALI show that there is no broad consensus regarding the right answer to the fair use and training model question. Courts are an excellent vehicle for resolving disputes between parties and interpreting laws. Courts are not built to craft broad policies regarding how GenAI will be built in the United States without harming authors, filmmakers, musicians, and other content creators. Ultimately, that is a job for legislators. If Congress does not act, then the administration may well try to mold copyright policy to its own goals. No matter the position Congress or the President take, they will find ample support and opposition in academia and industry. For now, watch closely to see who becomes the Register of Copyrights and how the ALI reacts to the recent resignations.

     

    Related Professionals

    -
    Andrew W. Coffman Andrew Coffman photograph

    Andrew W. Coffman

    Email

    Related Practices

    • Intellectual Property

    Related Industries

    • Artificial Intelligence (AI)
    Stay connectedReceive our latest thinking on topics you care about.SIGN UP NOW
    • ©2025 Phelps Dunbar LLP. All Rights Reserved
    • Lawyer Advertising
    • Privacy & Disclaimer
    • Contact Us
    © 2025 Phelps Dunbar LLP. All Rights Reserved