Competitive AI Use May Raise IP Issues

As more and more companies employ artificial intelligence (“AI”) in their business activities, novel legal questions continue to arise. Of particular note is the application of existing principles of intellectual property (“IP”) law to new AI technologies. Such concerns are not limited to enterprises in the content creation space. Given the manner in which AI algorithms are programmed, trained, and deployed, any company employing AI for competitive business purposes should be mindful of IP issues.

Recently, the United States District Court for the District of Delaware issued a ruling giving narrow breadth to the fair use doctrine in the context of AI technology. In Thomson Reuters Enterprise Centre GmbH & West Publishing Corp. v. Ross Intelligence Inc., the court applied existing copyright precedent arising from legacy technologies to find that the defendant’s use of the plaintiff’s content to generate a competing product using AI constituted copyright infringement.

The Decision

Thomson Reuters Enterprise Centre GmbH (“Thomson Reuters”) is a leading provider of legal research tools, providing the raw text of court decisions and other legal authorities as well as original content developed by Thomson Reuters itself. Thomson Reuters brought suit against Ross Intelligence Inc. (“Ross”), another legal research company offering an AI-assisted legal research service. The dispute arose from Ross’s use of copyrighted Thomson Reuters materials to train its AI algorithm.

Ross, a newcomer to the legal research field, originally sought to license Thomson Reuters’s materials to train its AI, but Thomson Reuters refused. Ross then made a deal with a third party, LegalEase, which provides legal research summaries called “Bulk Memos” that are derived from Thomson Reuters’s copyrighted materials. Ross used these “Bulk Memos” to train its AI, and Thomson Reuters sued for copyright infringement in 2020.

After several years of litigation, Thomson Reuters moved for summary judgment on certain of its claims for direct copyright infringement, and the parties filed cross-motions on Ross’s defense of fair use. Arising from principles of freedom of speech and codified in U.S. copyright statutes, fair use is an affirmative defense to copyright infringement that provides that a defendant is not liable for infringement if its copying of the plaintiff’s works is “fair” given the nature of the works at issue and the impact on the plaintiff.

The court ruled in Thomson Reuters’s favor as to certain of Thomson Reuter’s works, finding that Ross had directly infringed Thomson Reuters’s copyrighted materials and that Ross’s fair use defense failed as a matter of law. The court held that the Bulk Memos reproduced Thomson Reuters’s copyrighted material and that Ross’s use of the Bulk Memos to train its AI constituted direct copyright infringement.

Turning to Ross’s fair use defense, the court considered the four factors delineated in 17 U.S.C. § 107: (1) the use’s purpose and character, including whether it is commercial or nonprofit; (2) the copyrighted work’s nature; (3) how much of the work was used and how substantial a part it was relative to the copyrighted work’s whole; and (4) how Ross’s use affected the copyrighted work’s value or potential market.

The court held that the first factor favored Thomson Reuters because Ross’s use of the works was commercial (in direct competition with Thomson Reuters), and because Ross’s use of the works to train an AI was not transformative. Importantly, despite the technology involved in the case, the court rejected Ross’s argument that the use of the works to train AI was comparable to intermediate copying of computer code, instead emphasizing that the purpose of the copying was to produce a competing product. The court further held that Ross acted in bad faith by engaging in the copying after Thomson Reuters rebuffed its request for a license.

The court held that the second factor favored Ross because Thomson Reuters’s supplemental legal research materials are “not that creative.” However, the court noted that this factor rarely plays a significant role.

As to the third factor, the court held that this factor favored Ross because Ross’s end result was different from the copied works, and Ross did not make the actual Thomson Reuters works available to the public.

Finally, the court held that the fourth factor, which “is undoubtedly the single most important element of fair use,” favored Thomson Reuters. The court emphasized again that Ross was using Thomson Reuters’s copyrighted materials to develop a competing product, which made Ross’s fair use defense untenable.

Accordingly, the balance of the factors defeated Ross’s fair use defense, and the court granted summary judgment in Thomson Reuters’s favor as to copyright infringement.

Practical Considerations

The Thomson Reuters decision is an enlightening early example of how courts may deal with novel questions concerning AI technology. As is typical in copyright cases involving new technology, the court applied existing precedent involving legacy technologies. The court emphasized facts such as actual copying and direct competition to find a lack of fair use. While the court acknowledged that the copyrighted works were manipulated in new ways through AI technology, the end result—a product based on a copyrighted work competing with the rights holder—appeared to be the deciding factor.

In view of this, companies should be thoughtful about the use of AI in developing new products, particularly if AI algorithms are being used to analyze or summarize information generated by competitors. The mere fact that a technology is new has seldom succeeded a defense to copyright infringement.

Finally, it should be noted that this decision is subject to appeal, and that numerous other cases testing the intersection of AI and IP are currently pending. The law will continue to develop.

If you have additional questions related to the fair use doctrine in the context of AI technology, reach out to your Quarles attorney or:

• Matthew Holohan: (303) 381-9891 / matthew.holohan@quarles.com