In Re: OpenAI, Inc. Copyright Infringement Litigation

MDL 3143 consolidates multiple copyright infringement actions against OpenAI and Microsoft before Judge Sidney Stein in the U.S. District Court for the Southern District of New York. The coordinated proceedings bring together a broad coalition of authors, news organisations, and digital content creators who allege that OpenAI's large language models, including ChatGPT, were trained on copyrighted text and audiovisual material without consent or compensation.

The Basbanes Plaintiffs (Nicholas A. Basbanes and Nicholas Gage) [filed 5 January 2024] raise similar claims focused on the wholesale ingestion of literary works. The Authors Guild Plaintiffs (The Authors Guild and 28 prominent writers including George R.R. Martin, John Grisham, and Michael Chabon) [filed 19 September 2023] contend that OpenAI copied and reproduced entire books to develop its AI systems. The Tremblay Plaintiffs (Paul Tremblay, Sarah Silverman, and others) [filed 28 June and 7 July 2023, N.D. Cal.] allege direct and vicarious copyright infringement, asserting that ChatGPT can generate derivative works based on their writing.

The News Media Plaintiffs (The New York Times [filed 27 December 2023], Daily News [filed 30 April 2024], Center for Investigative Reporting [filed 27 June 2024], The Intercept, Raw Story, and AlterNet [each filed 28 February 2024]) allege misappropriation of journalistic content to train OpenAI's models. The Millette Plaintiffs, led by YouTube creator David Millette [filed 2 August 2024, N.D. Cal.], represent video creators whose content was allegedly transcribed without consent for AI training.

The Sullivan Plaintiffs (Michael Sullivan, Kenneth Saladin, Zvi Bodie, Alan Marcus, Kevin Patton, Lion Den Inc., and Dana Loewy) [filed 14 August 2026] bring a putative class action on behalf of textbook authors, alleging that OpenAI sourced their works from pirated "shadow libraries" such as LibGen, stripped copyright management information using a purpose-built script, and trained GPT-N on them. They argue that the adoption-driven textbook market is uniquely vulnerable to substitution by LLM outputs, and name Microsoft as vicariously and contributorily liable for providing the Azure infrastructure on which the copying occurred. wikiHow, Inc. [filed 21 August 2026] asserts 1,211 registered copyrights covering 11,211 how-to articles and pleads infringement at three stages: ingestion for training, retrieval-augmented generation at runtime, and verbatim reproduction in ChatGPT outputs. It also alleges DMCA violations for the removal of bylines and copyright notices via extraction tools such as Dragnet and Newspaper, and points to continued crawling in breach of its robots.txt directives and terms of use after failed licensing talks.

Microsoft is named as a co-defendant for its financial backing, cloud infrastructure, and commercial integration of OpenAI's models, which plaintiffs argue make it jointly liable for the alleged infringement.

The central question across the MDL is whether the use of copyrighted material for AI training constitutes fair use or direct infringement.

On 1 September 2026, the United States filed a Statement of Interest under 28 U.S.C. § 517, applicable to all matters in the MDL, taking the position that AI training is an "extraordinarily transformative" fair use purpose and describing plaintiffs' market-dilution theory as "deeply flawed". The Department of Justice also warned that requiring licences would hand the biggest technology firms a near-monopoly on model building.

This summary is provided for educational and research purposes only and does not constitute legal advice. Allegations described are those made in the filings. Please consult the underlying source documents.