The U.S. Court of Appeals for the Third Circuit rejected an artificial intelligence company’s argument that using copyrighted material for training purposes was fair use. In Thomson Reuters v. Ross, the court ruled for Thomson Reuters in a copyright infringement case against Ross Intelligence.
This is the first federal appellate decision directly addressing copyright fair use in the AI-training context, according to reports. The Third Circuit upheld the court’s original ruling that Thomson Reuters headnotes have enough original content to warrant copyright protection.
Headnotes are short descriptions of legal issues that appear before judicial opinions. Ross used a third-party tool to glean information from a Reuters-owned legal research platform called Westlaw. It then used that content and data to train its own, similar legal tool, Courthouse News Service reported.
Fair Use
Fair use under the US Copyright Act of 1976 allows copyrighted materials to be used for purposes such as criticism, journalism, and research. Fair use is the cornerstone of AI firms’ argument that they’re not wrong to use anyone’s material to train their new, innovative technology.
Following months of deliberation, U.S. Circuit Judge Tamika Montgomery-Reeves wrote for the panel on Wednesday, Oct. 7: “Under ROSS’ framing, this case appears to concern the future of AI legal technology. But appearances can be deceiving. In truth, this is no more than an ordinary copyright case.”
Ross no longer exists, and its AI wasn’t generative; it’s not clear how this win will affect other cases. The Copyright Lately blog said the win is a win-win for everyone except Ross, with everyone claiming victory. Except Ross, whose people are taking this to the highest court, because defending the lawsuit put it out of business. In 2021. “That makes Ross something like the Sony Betamax of AI copyright …” said CopyrightLately, which explains all of it.
Yar Chaikovsky of White & Case LLP told Courthouse News, “We intend to seek review by the Supreme Court to obtain much-needed clarity on these issues, which carry significant implications for innovation and the development of AI technologies.”
More Lawsuits
Over 100 copyright-infringement lawsuits are currently in play against many AI companies. The music industry, newspaper and media firms, comedians, and other creatives have filed individual and group lawsuits. Many of these cases raise fundamental questions about whether training AI on copyrighted works constitutes fair use, and whether AI-generated outputs infringe copyrights.
A copyright case expected to be heard this month is Concord Music Group versus Anthropic PBC, in the U.S. District Court for the Northern District of California. Several music publishers brought the lawsuit, accusing AI developer Anthropic of illegally using copyrighted lyrics and musical works to train its AI model, Claude.
Over the summer, a federal judge approved a landmark $1.5 billion class-action settlement between Anthropic and thousands of authors and publishers. While the court previously ruled that training language models on books can constitute fair use, it found that Anthropic’s storage of millions of unauthorized books in a central pirated library violated copyright law.
