We have been closely following, and have written about, the ongoing cases regarding AI companies using copyrighted materials to train their AI models. See: https://sharmalawpllc.com/the-new-york-times-challenges-openai-and-microsoft-over-the-use-of-its-content-to-train-ai-chatbots-in-openai-copyright-battle/ and https://sharmalawpllc.com/major-record-labels-sue-ai-music-generating-software/. Most of these cases revolve around a similar fact pattern: a company uses copyrighted works to train AI models without the copyright holder’s permission. The authors of the copyrighted material argue that having their works used without their permission to train AI models constitutes copyright infringement. On the other side, the AI companies and developers argue that their use of the copyrighted material to train their AI models is protected by “fair use”, and therefore does not infringe the copyright of the authors. Simply put, “fair use” is an affirmative defense that can be raised in response to claims by a copyright owner that a person is infringing a copyright. Fair use permits a party to use a copyrighted work without the copyright owner’s permission for purposes such as criticism, comment, news reporting, teaching, scholarship, or research. Whether or not using a copyrighted work without the copyright owner’s permission constitutes fair use is decided on a case by case basis, and many factors are considered, as discussed in more detail below.

The first such case to have reached a conclusion on this issue was recently decided in the U.S. District Court for the District of Delaware (Thomson Reuters Enterprise Centre GmbH & West Publishing Corp. v. ROSS Intelligence Inc.). In this case, Ross Intelligence Inc. (“Ross”) was using copyrighted material from Thompson Reuters (“Reuters”). Reuters owns and operates a law platform called “Westlaw”, which has case law opinions for legal research purposes. In Westlaw, there are headnotes, which have editorial content created by Reuters that summarizes the results of the legal cases. Ross was using these headnotes to train their own AI legal research platform which would compete with Westlaw. It was ruled in this case that the Westlaw headnotes constituted copyrighted material, and that Ross had infringed Reuters’ copyright by using the headnotes in this manner.

Ross tried to argue that its use of the headnotes was fair use by claiming that their use was “transformative” (i.e. that their use added something new, with a different purpose and character, and did not substitute for the original work). In analyzing a fair use defense, a court will consider four factors: (1) the purpose and character of the use (e.g. commercial vs. educational), (2) the nature of the copyrighted work, (3) the amount and substantiality of the portion used, and (4) the effect on the potential market for or value of the copyrighted work (e.g. does it compete with/replace the copyrighted material?). The court in this case found that the purpose and character of Ross’s use was solely commercial and their use did not have a “further purpose or different character”. The court also found that Ross’s platform was being created to directly compete with Westlaw, and would in fact be a potential market substitute.

It is notable that the Reuters case did not involve generative AI, where many of the other cases currently before the courts do. Ross’s platform was not designed to create an original output based on a user’s input using the AI model’s training; rather it was to be a tool that used AI to search legal opinions that resided in its database. This is an important distinction, as defendants in other suits involving generative AI will likely continue to argue that the purpose of their use of copyrighted content is to create an entirely new work, which does not compete with the underlying copyrighted work. We are keeping a close eye on the cases that are still moving through the court systems across the U.S. and will keep you informed of new developments.

If you need assistance in the area of AI and/or copyright infringement, please contact our attorneys here.