Law360 (September 4, 2026, 6:20 PM EDT) -- Microsoft on Friday told a Manhattan federal court that using books and news articles to train the large language model behind the artificial intelligence model Copilot counts as fair use, and the company should therefore be granted a favorable judgment.

Microsoft said in a pair of motions that it should be granted summary judgment on claims brought by authors and news organizations that their works were illegally ingested to train Copilot.

Microsoft opened one brief by quoting the mathematician Ada Lovelace, who postulated a "poetical science," contrasting with the theories of her parents — poet Lord Byron and educational reformer Annabella — who believed literature and science were at odds. Lovelace saw in early computers an "engine of progress that could wield any complex system humans could devise," Microsoft said.

"Ada was right. What better proof of poetical science than the advance at issue in this case: the Large Language Model, a computer system that harnesses the deep structure of language and knowledge to give people a profound tool for speech, communication, and learning," the motion said.

Copyright law defends the public's right to build upon previous works via fair use principles, which is why search engines and certain phone applications are legal, the motion said. Those technologies resulted in rights owners seeking to bar them, but "each time the law said otherwise," according to the motion.

The news organizations, which include The New York Times, Center for Investigative Reporting and several regional newspapers, do not seriously contest that LLMs are transformative but instead seek to deflect by pointing to OpenAI's alleged pirating of some books, Microsoft said.

Each of the four fair use factors "dramatically" favors a finding that AI training is transformative, Microsoft said. LLMs have already been used for expressive uses that "copyright law cherishes," including assisting attorneys, scientists and doctors, the motion reads.

AI comes with detractors like any other new technology and will require care in development, but a copyright infringement case is not the venue to litigate every policy implication that comes with it, Microsoft said.

Microsoft said the fair use factors favor it as to the ingested news articles as well. The company said it "has deep respect for the critical role News Plaintiffs have long served in our public discourse," but that did not allow news organizations to "block transformative new technologies that do not replace their protected expression."

The alleged market dilution asserted by the news organizations was not cognizable, and the concept turns copyright upside down, according to Microsoft. The creation of noninfringing news content that competes with the news organizations are "copyright goods, not evils," the motion said.

Microsoft said Copilot's grounding feature, in which it scans the internet for answers to queries instead of relying on training data, was fair use as a matter of law as well. Web grounding is a "transformative new product much different" from the asserted news content, and the court should hold that any outputs that incidentally resemble news content count as fair use, the motion said.

In its own summary judgment motion, OpenAI said Friday that the undisputed facts show that OpenAI's purpose in pretraining ChatGPT was to derive broad statistical patterns related to language that can be used to create new text instead of making use of protected copyrighted works. That purpose is totally different from the purpose for which the authors wrote their books, and the authors and news organizations have not been harmed by OpenAI's conduct, the motion said.

OpenAI said ChatGPT does not display copies of books and that a tiny fraction of its logs regurgitated content from those books, typically only amounting to a few dozen words.

The news organizations also filed a motion for summary judgment Friday, seeking a favorable ruling of liability at "each stage of the AI pipeline," including the acquisition of articles, training the models on journalistic content, grounding via the web and outputting responses to user inquiries.

The news organizations said there was no precedent for Microsoft and OpenAI's claims that unauthorized copying of millions of news articles counted as fair use. The purpose and character of the tech companies' use was to "exploit the works for their intrinsic value" and offer substitute products that harm the market for original journalism, the motion said.

"The documents and testimony uncovered during discovery lead to only one conclusion: Microsoft and OpenAI stole from The New York Times to make commercial products that substitute for its journalism, threaten its business, and undermine its industry. We look forward to Microsoft and OpenAI being held accountable for their theft," the Times' attorney Ian Crosby of Susman Godfrey said in a statement.

Asked for comment Friday, an OpenAI spokesperson referred to a statement on its website titled "reporting the facts about the New York Times' lawsuit" that said, in part, "Fair use benefits everyone. The New York Times has tried to present this lawsuit as a binary choice between human authorship and AI. That's simply wrong. It's also inconsistent with even their own use of this technology, not to mention the experience of many others within the news and publishing industries who are using AI in ways that keep human authorship front and center."

A representative for Microsoft declined to comment. An Authors Guild representative also declined to comment.

The Times is represented by Ian Crosby, Katherine Peaslee, Davida Brook, Emily Cronin, Adnan Muttalib, Elisha Barron, Zachary Savage, Tamar Lusztig, Alexander Frawley and Eudokia Spanos of Susman Godfrey LLP and Steven Lieberman, Jennifer Maisel and Kristen Logan of Rothwell Figg Ernst & Manbeck PC.

The regional newspapers are represented by Steven Lieberman, Jennifer Maisel, Kristen Logan, Robert Parker, Jenny Colgate, Mark Rawls, Michael Jones, Alexandra Hughes, Bryan Thompson and Mary Mullins of Rothwell Figg Ernst & Manbeck PC.

The Center for Investigative Reporting and The Intercept Media are represented by Jon Loevy, Michael Kanovitz, Matthew Topic, Thomas Kayes, Steven Art, Kyle Wallenberg, Shelley Geiszler, Stephen Stich Match and Lauren Sonnenberg of Loevy & Loevy.

OpenAI is represented by Allison S. Blanco, Sarang V. Damle, Luke A. Budiardjo, Andrew M. Gass and Elana Nightingale Dawson of Latham & Watkins LLP, Joseph C. Gratz, Tiffany Cheung and Caitlin Sinclaire Blythe of Morrison Foerster LLP and Michelle S. Ybarra, Robert A. Van Nest, R. James Slaughter, Nicholas Goldberg and Paven Malhotra of Keker Van Nest & Peters LLP.

Microsoft is represented by Jared B. Briant, Kirstin L. Stoll-DeBell, Carrie A. Beyer, Jeffrey S. Jacobson and Elizabeth M.C. Scheibel of Faegre Drinker Biddle & Reath LLP and Annette L. Hurst, Christopher Cariello, Marc Shapiro, Sheryl Koval Garko and Laura Brooks Najemy of Orrick Herrington & Sutcliffe LLP.

The case is In re: OpenAI Inc. Copyright Infringement Litigation, case number 1:25-md-03143, in the U.S. District Court for the Southern District of New York.

--Editing by Vaqas Asghar.

For a reprint of this article, please contact reprints@law360.com.