Written by: Juliana Cipolla

On December 27, 2023, The New York Times (“NYT”) filed a complaint against Microsoft and OpenAI claiming defendants allegedly created generative AI products relying on large-language models that were built by copying and using millions of NYT’s copyrighted works including news articles, investigations, opinion pieces, and more. The NYT emphasizes its exclusive rights of reproduction, adaptation, publication, performance, and display held under copyright law. The NYT further alleges that both Microsoft and OpenAI’s copying of NYT’ works maintains user direct engagement with Defendants’ own websites and applications instead of referring users to the source of information, namely the NYT.
Microsoft and OpenAI want the court to throw out NYTs’ claims of infringement involving older NYT articles and the Defendants’ misappropriation of copyrighted works. So, what’s left? The heart of the lawsuit, whether the use of copyrighted work scraped from the internet to train and test the large language models central to generative AI is protected by the Fair Use doctrine in Copyright to exempt the defendants of getting permission for the source information beforehand.

The Times provided dozens of examples of Open AI’s ChatGPT answering prompts allegedly sourced with large amounts of information that mimics the NYT newspaper’s stories. OpenAI, in its motion to dismiss, accused NYT of hiring a software consultant to forensically obtain dozens of examples showing the allegedly infringed copying of NYT’s copyrighted works. OpenAI alleges that NYT consultant identified and exploited a bug through a hack by feeding large amounts of article texts into a chat prompt to induce the tool to respond by regurgitating long passages of text, known as unintended memorization.

This is not the first time a court has looked at the copyright of a work generated by artificial intelligence. In 2023, Thaler v. Perlmutter, issued by the D.C. Circuit Court, opined on the issue whether a copyright should be issued and transferred to the machine’s owner. The court ultimately concluded that artwork generated by artificial intelligence cannot be copyrighted.

This case has spurred questions about whether there should be rules relating to ownership of works that come out of artificial intelligence. This case could play a major role in how the NYT’s case is to be handled. There are three main issues that will incidentally be discussed. First, is there a copyright that can be attached to the work that derived by an AI system. Second, are there copyright properties already attached to AI generated works containing mimicked texts of copyrighted work. And third, will the AI generated mimicked texts of previously copyrighted works be deemed property of the NYT, or Microsoft and OpenAI.

The Beckage Firm has deep experience in AI topics, litigation, putative class actions, and compliance matters.