Bloomberg asks US court to toss copyright lawsuit over AI training

March 25 (Reuters) - Bloomberg LP has asked a New York federal judge to dismiss a lawsuit from Arkansas governor Mike Huckabee and other authors who claimed the company misused their books to train its large language model BloombergGPT.

"Bloomberg's alleged 'use' of Plaintiffs' copyrighted works as part of a research project into the capabilities of generative AI falls squarely within the ambit of the fair use doctrine," the company said in a court filing on Friday.

Representatives for the authors and for financial data and news company Bloomberg did not immediately respond to requests for comment on the filing on Monday. Bloomberg and Reuters parent Thomson Reuters are competitors.

Huckabee, best-selling Christian author Lysa TerKeurst and other writers sued Bloomberg, Meta Platforms META.O and Microsoft MSFT.O last year, accusing them of using a dataset that contains thousands of pirated books to teach text-based artificial intelligence systems how to respond to human prompts.

The lawsuit is one of several major cases brought by copyright holders against tech companies over the alleged misuse of their content to train generative AI. The writers' claims against Meta and Microsoft have since been separated from the Bloomberg case and moved to California.

Huckabee, TerKeurst, and authors David Kinnaman, Tsh Oxenreider and John Blase told the court that their books were used without permission to train BloombergGPT, a generative artificial-intelligence system for financial analysis.

Bloomberg responded on Friday that the authors' claims lacked specific allegations about how it infringed or which of their books were misused for BloombergGPT, which it called an internal company application.

The company told the court that it made fair use of the authors' works, a doctrine of U.S. copyright law that allows for the unauthorized use of copyrighted material in some circumstances. Fair use is expected to be a key defense for tech companies in copyright disputes over AI training.

"Simply put, a limited and private use of copyrighted works by a news reporting enterprise to teach a not-for-commercial-use AI model as part of an internal research project into the capabilities of generative AI, is not copyright infringement," Bloomberg said.

The case is Huckabee v. Meta Platforms Inc, U.S. District Court for the Southern District of New York, No. 1:23-cv-09152.