Copyright claims against AI companies has just gained a possible boost.
US Federal Judge last week delivered a summary judgment In a case brought by TE Technic Conglomerate Thomson Reuters against legal technology firm Ross Intelligence. The judge found that Ross’s use of Reuters ‘content to train its AI law research platform violated on Reuters’ intellectual property.
The result could have implications for the more than 39 Copyright AI processes Currently working with US courts. This is said, it is not necessarily a SLAM dam for plaintiffs who claim that AI companies have violated their IP rights.
All about the headnotes
Ross has been charged with using Headnotes – summaries of legal decisions – from Westlaw, the legal research service of Reuters, to train its AI. Ross has marketed his AI as a tool to analyze documents and perform question-based searches through court recordings.
Ross argued that its use of copyrighted head notes was legally defensible because it was transformative, which means it repulsed the head notes to serve a significantly different function or market. In his summary judgment, Stephanos Bibas, the judge presiding over the case, did not find that argument particularly compelling.
Ross, Bibas said in his opinion, regained Westlaw Headnotes in a way that directly replied Westlaw’s legal research service. The runway platform has not added a new meaning, purpose or comment, Bibas determined – undermining Ross’s claim of transformative use.
In his decision, Bibas also cited Ross’s business motivations as a reason that the defense from the beginning missed the mark. Ross sought to benefit from a product that competed directly with Westlaw, and without a significant “recontectualization” of the IP-protected Westlaw material.
Shubha Ghosh, a professor at Syracuse University, who studies IP law, called it a “strong victory” for Thomson Reuters.
“The process will continue, [but] Thomson Reuters was awarded a summary judgment, a victory at this stage of the trial, “Ghosh said.” The judge also claimed that Ross was not allowed to summarize a judgment on his defenses, such as fair use and merging. As a consequence, the case continues to judge with a strong win for Thomson Reuters. ”
Narrow in application
Already, at least one set of plaintiffs in another AI -copyright case has asked a court to consider Bibas’ decision. But it is still unclear whether the precedent will swing other judges.
Bibas’s opinion made a point to distinguish between “generative AI” and the AI, which Ross used, which did not generate content, but only spit judicial opinions that were already written.
Generative AI, which is in the center of copyright processes against companies such as Openai And MidjourneyIt is often trained on massive amounts of content from public sources around the site. When feeding many examples, generatively AI can generate speech, text, pictures, videos, music and more.
Most companies developing generative AI argues that Fair use -doctrines Shield their practice scrap data and use it for training without offset – or even credit – the owners of the data. They argue that they have the right to use any publicly available content for training and that their models actually result in transformative works.
But not every copyright owner agrees. Some point to the phenomenon known as Regurgitationwhere generative AI creates content closely similar to the work on which it was trained.
Randy McCarthy, a US patent lawyer at Hall Estill’s lawyer, said Bibas’ focus on the “effects on the original work market” could be key to the cases of owners of rights against generic AI developers. But he also warned that Bibas’s opinion is relatively narrow and that it can be overturned on appeal.
“One thing is clear, at least in this case: just use copyright material as training data [for] It cannot be said that AI is not a fair use, “McCarthy told Techcrunch.”[But it’s] One battle in a larger war, and we will need to see more developments before we can extract from this law on the use of copyrighted materials such as AI -training data. ”
Another lawyer Techcrunch spoke with, Mark Lezama, a lawsuit partner at Knobbe Martens focusing on patent disputes, believes that Bibas’s opinion could have broader implications. He believes that the judge’s reasoning could extend to generative AI in his various forms.
“The court rejected fair use defense as a law partly because Ross used [Thomson Reuters] Headnotes to develop a competitive legal research system, “he said.” Although the court alluded to that it might be different from a situation involving generatively AI, it is easy to see a news site arguing that copying their articles to training generatively AI is not different because the generative AI uses The copyright articles to compete with the news site for user attention. ”
In other words, publishers and owners of copyright come up with it with AI companies have small reasons to be optimistic after the decision – emphasis on slight.
