52 Friend-of-the-Court Briefs Land in the OpenAI Copyright Fight

The court overseeing the copyright lawsuits against OpenAI has received 52 amicus briefs, and the pile of paper is growing as the case moves toward a hearing. An amicus brief, from the Latin for "friend of the court," is a written argument filed by someone who is not a party to the case but wants the judge to hear their view. The deadline was October 30, and Publishing Perspectives reports that 26 of the briefs back the plaintiffs, 18 back the defendants, and 8 take neither side.
The case, In re OpenAI Copyright Infringement Litigation, brings more than a dozen separate lawsuits together so one judge, Sidney H. Stein, can hear them as a single matter. The plaintiffs include the Authors Guild, a US writers' organisation, and news organisations including the New York Times. Microsoft is a co-defendant alongside OpenAI.
The central argument is about how AI chatbots learn. The plaintiffs say OpenAI copied books and other works to train its models, including from pirated sources, and that the answers the AI produces can also infringe copyright. OpenAI says copying for training is fair use, a legal exception that allows some use of protected material without permission, and that its outputs do not infringe.
Publishers have joined in. Four publishing associations, including the Association of American Publishers and the Independent Book Publishers Association, backed the Authors Guild and other groups in a filing on October 29. The Association of American Publishers describes the conduct at issue as sweeping, unauthorized use of copyrighted works. Authors Alliance, a group that supports authors' access to their own work, asked the court for permission to file on October 29 and then filed in support of the defendants on October 30.
The US government has also taken a position. On September 4, the Department of Justice filed what is called a Statement of Interest, a document in which the government tells a court its view, urging the judge to find that using copyrighted works to train AI is fair use.
Not every filing has gone the way one side wanted. Judge Stein struck from the evidence a study that suggested AI-written books were displacing human-written ones. OpenAI had argued that the study's author did not disclose a funding conflict, and the judge agreed to remove it.
The next step is procedural but important. A final round of briefs is due on November 6. After that, oral arguments, where lawyers make their case in person before the judge, are expected to be scheduled. No date for them has been announced.
For readers, the case sits at an odd crossroads. A book is written by a person, copied into a machine, and now debated in a courtroom by lawyers, writers, publishers, and the government. Whatever the judge decides, the question of who gets paid when words are used to build a machine will not go away quietly.


