Thomson Reuters Had the First Massive Win in an AI Copyright Case

• Thomson Reuters scored an early victory in an AI-related copyright case in opposition to Ross Intelligence.
• The ruling highlights honest use limits. Honest use is on the heart of ongoing AI copyright lawsuits.
• Authorized specialists say the case differs from different litigation involving generative AI companies.

Content material and expertise conglomerate Thomson Reuters this week scored the primary huge win in a US synthetic intelligence-related copyright case.

A federal decide’s Delaware ruling in favor of Thomson Reuters on the authorized doctrine of “honest use,” nonetheless, doesn’t imply that the slew of authors and publishers who’ve sued generative AI firms like OpenAI for copyright infringement can count on to have the identical sort of success, tech legislation specialists instructed Enterprise Insider.

The ruling, although, may have an affect on the end result of these circumstances which are presently winding by means of the courts, one of many tech legislation specialists stated.

Thomson Reuters sued the now-shuttered authorized AI startup Ross Intelligence in 2020, arguing that Ross infringed its copyrights when the startup used content material from Thomson Reuters’ Westlaw authorized analysis database to create a competing platform that makes use of synthetic intelligence.

In his revised ruling Tuesday, US District Court docket Choose Stephanos Bibas shot down Ross’ honest use protection and as an alternative granted a abstract judgment for Thomas Reuters on honest use.

“None of Ross’s doable defenses holds water. I reject all of them,” Bibas wrote within the ruling, during which he defined that in his 2023 opinion within the case, he denied abstract judgment on honest use.

The query of honest use is on the coronary heart of main ongoing copyright lawsuits in opposition to generative AI companies. Firms like OpenAI have pointed to the honest use exemption to copyright legal guidelines with a purpose to justify their use of copyrighted materials to coach AI fashions.

Authorized specialists instructed BI that there are key nuances in Thomson Reuters’ case in opposition to Ross and the blockbuster litigation involving generative AI companies — one being that the Thomson Reuters case would not must do with generative AI expertise.

Bibas made that distinction in his ruling, writing, “Ross was utilizing Thomson Reuters’s headnotes as AI information to create a authorized analysis software to compete with Westlaw. It’s undisputed that Ross’s AI shouldn’t be generative AI (AI that writes new content material itself). Fairly, when a person enters a authorized query, Ross spits again related judicial opinions which have already been written.”

Mark Bartholomew, a College at Buffalo legislation professor, instructed BI that he doesn’t assume the ruling may have a dramatic impact on the opposite main AI-related copyright circumstances because it’s only one opinion of a decrease courtroom and “neglects among the most vital honest use case legislation.”

“Nonetheless, the plaintiffs in these upcoming circumstances, like The New York Instances, must be proud of the end result right here,” Bartholomew stated, including that these plaintiffs will possible “trumpet this determination of their authorized briefs.”

The New York Instances sued OpenAI for copyright infringement in 2023. The New York Instances declined to remark for this story. OpenAI didn’t instantly reply to a request for remark.

One of many different ways in which the Thomson Reuters case is totally different from different ongoing AI-related copyright infringement litigation is that Ross Intelligence was a direct competitor of Thomson Reuters, Bartholomew identified.

“In most of the different copyright AI circumstances to be determined, the defendant is accused of copying to coach the AI for an arguably new goal that doesn’t instantly compete with the unique,” Bartholomew stated. “That will enable this case to be distinguished, permitting the AI platforms to argue they nonetheless fulfill the honest use protection whatever the holding in Thomson.”

Harry Surden, a professor on the College of Colorado Regulation Faculty, stated that he fears judges overseeing different AI copyright lawsuits could not keep in mind the variations between the circumstances.

The ruling is “prone to confuse judges and others on the difficulty of generative AI, however there are vital variations which are very refined and most of the people will miss,” Surden stated, explaining, “It is each a unique expertise and the corporate was appearing sort of duplicitously right here.”

For these causes, Surden stated the ruling shouldn’t have a lot affect on the continuing AI copyright circumstances. He thinks it’ll anyway.

“Judges have a tendency to not be specialists in copyright legislation or generative AI, so I believe they’re prone to miss a variety of these nuances, and I believe it’ll, sadly, impact different circumstances,” stated Surden.

James Gatto, a companion on the legislation agency Sheppard Mullin who co-leads the agency’s AI trade staff, instructed BI it stays to be seen what sort of ripple impact, if any, the ruling in favor of Thomson Reuters may have.

“Honest use is a fact-specific inquiry distinctive to every case,” stated Gatto.

Although the decide’s determination doesn’t tackle generative AI, Gatto stated it “reinforces the bounds of honest use, significantly in circumstances the place copyrighted materials is used for non-transformative functions to develop a competing product.”

“The courtroom simply decided that Ross’s use was business, as Ross sought to revenue from the copyrighted materials with out paying the customary value,” stated Gatto.

In the meantime, Thomson Reuters hailed the authorized win, saying in a press release to BI: “We’re happy that the courtroom granted abstract judgment in our favor and concluded that Westlaw’s editorial content material created and maintained by our legal professional editors, is protected by copyright and can’t be used with out our consent. The copying of our content material was not ‘honest use.'”

Representatives for Ross didn’t instantly reply to a request for remark.