The court recently decided that a previous rival of Thomson Reuters is prohibited from utilizing the company’s content to establish an AI-powered legal platform, a ruling that may serve as a precedent for future cases concerning the lawfulness of employing copyrighted materials to educate AI systems.
In a court decision made on Tuesday, U.S. District Judge Stephanos Bibas declined Ross Intelligence’s claims that their actions fell under the “fair use” exemption regarding copyright laws. This judgment on an unprecedented matter could serve as a precedent for creators who are taking legal action against tech companies in the entertainment industry, even though this specific case doesn’t concern the development of AI-generated content.
“Originality is central to copyright,” Bibas wrote.
Immediately following the court’s decision, Concord Music Group requested that the federal judge presiding over their lawsuit against Amazon-owned Anthropic, regarding the utilization of song lyrics to train Claude, take the verdict into account when assessing the overall case.
The main focus of this matter concerns the Thomson Reuters platform for legal research, where users purchase access to a variety of information such as court cases, state and federal laws, law reviews, and regulations. This material includes headnotes that provide concise summaries of legal points and case outcomes, which are protected by copyright.
Previously operating AI company Ross, supported by venture firm Y Combinator, employed similar notes to those used in training a rival legal search engine when Thomson Reuters refused to grant a license for the content. The distinct feature of this lawsuit compared to others involving AI is that an intermediary re-purposed copyrighted work for AI instruction. In contrast, creators accuse OpenAI, Meta, Anthropic, and other entities of outright plagiarism in their lawsuits.
In the court decision made on Tuesday, Judge Bibas indicated that there might be potential violations of over 2,200 sections by Ross. The jury will assess compensation based on whether any copyrights owned by Thomson Reuters have lapsed.
The court’s verdict hinged on whether the headnotes could be considered as unique works eligible for protection under copyright law. Judge Bibas, who previously ruled in favor of Ross Intelligence during summary judgment in 2023 (a decision that was later retracted prior to trial), supported Thomson Reuters’ stance on this matter. According to him, headnotes can demonstrate creativity by summarizing, synthesizing, or interpreting parts of an opinion.
In addition, each headnote stands independently as its own copyrightable piece,” the judge stated. “This became apparent when I compared the lawyer’s editorial decisions to those of an artist. Just as a raw block of marble or a judicial opinion itself isn’t protected by copyright, so too is a sculptor’s work not initially. However, it’s the sculptor’s choice of what to remove and keep that gives rise to the copyrighted sculpture.
Note that Bibas did not rule in favor of fair use, which allows the use of copyrighted material to create a new work, as long as it’s transformative. In this case, Bibas pointed out that Ross aimed to generate profits by using Thomson Reuters headnotes, which often leans against fair use claims. The court emphasized that even if we consider Ross favorably, their intention was to compete with Westlaw by creating a rival product. Whether Thomson Reuters used the data to train its own legal search tools is irrelevant; the potential impact on AI training data market is sufficient reason for not granting fair use.
The court frequently referred to the Supreme Court’s decision in the case of Andy Warhol Foundation for the Visual Arts v. Goldsmith, which significantly limited fair use. In this instance, the majority stated that the evaluation of whether an infringing work has been sufficiently transformed should be weighed against its “commercial nature.” Artists are using this ruling to argue that AI companies could have obtained a license for the copyrighted material instead, and their markets were negatively impacted as a result.
Randy McCarthy, a legal expert in intellectual property at Hall Estill, states that the court’s decision will be widely celebrated by artist and content creator communities as a significant step forward in their lawsuits against other generative AI systems. He further notes that simply using copyrighted material to train an AI should not automatically be considered fair use.
The main points of contention for widespread acceptance of AI revolve around whether existing legal principles will be applicable.
Read More
- ‘Violent Night 2’ Writers Discuss Who Could Star as Mrs. Claus
- Jahan Loh and Stash Team Up for ‘Namecaps: Spaceman’ Exhibition at Central Chidlom
- ‘Fast & Furious’ Feud Continues After Vin Diesel’s “Petty” Gesture at Golden Globes
- Azaad actress Rasha Thadani reacts to comparison with Suhana Khan, Janhvi Kapoor, Khushi Kapoor; ‘They have already…’
- Sky Force actor Veer Pahariya gifts his locket to paparazzo; debutant’s sweet gesture goes viral: WATCH
- ANTi COUNTRY CLUB TOKYO’s Ode to American Prep
- Supreme Court to Review Nvidia’s Crypto-Mining Revenue Case
- AMBUSH Delves Into Its Archives for New Book With Rizzoli
- Square Enix has a Punishment for Toxic Fans Who Harass Their Employees
- Damon Wayans on Why He “Purposefully” Got Himself Fired From ‘SNL’: “I Snapped”
2025-02-12 04:25