New York Times-led group asks court to sanction OpenAI in US copyright dispute

By Blake Brittain

July 9 (Reuters) - A radical of newspapers including the New York Times and New York Daily News asked a national tribunal successful Manhattan ‌on Thursday to authorisation OpenAI successful their high-stakes copyright quality for allegedly ‌lying to the tribunal astir its quality to hunt its systems for impervious that it misused millions of ​their articles successful AI training.

The newspapers told the tribunal successful a filing that OpenAI falsely told the tribunal it could not hunt its ample connection models for their copyrighted worldly portion hiding that it had done truthful "even earlier the archetypal News Plaintiff filed suit."

The newspapers ‌said that OpenAI had besides ⁠deleted billions of applicable ChatGPT conversations oregon made them unsearchable. They asked the tribunal for sanctions, including attorneys' fees, and a tribunal uncovering ⁠that OpenAI's chat logs showed that the institution misused their copyrighted works.

Spokespeople for OpenAI did not instantly respond to a petition for remark connected the motion.

The lawsuit, archetypal filed by the ​Times successful ​2023, accused OpenAI and its largest fiscal ​backer Microsoft of utilizing millions of ‌its articles without support to bid the ample connection exemplary down OpenAI's fashionable chatbot ChatGPT.

The lawsuit is 1 of galore brought by copyright owners including authors, ocular artists and euphony labels against tech companies specified arsenic OpenAI, Anthropic and Meta Platforms for allegedly misusing their worldly to bid AI systems.

"For implicit 2 years, OpenAI lied to The Times, ‌The Daily News Plaintiffs, the public, and the ​court," the New York Times' pb lawyer Ian Crosby ​said successful a statement. "It claimed searching ​ChatGPT outputs for copies of The Times' and the Daily News ‌Plaintiffs' contented was infeasible, burdensome, and invasive ​of users' privateness – portion ​at the aforesaid clip concealing that it had already done specified searches."

OpenAI antecedently told the tribunal that it did not person tools to hunt its datasets ​and output logs for copyrighted ‌material, but an OpenAI worker aboriginal testified that the institution had "performed aggregate ​searches for News Plaintiffs' content," according to the newspapers' Thursday filing.

(Reporting by Blake ​Brittain successful Washington; edited by David Gaffen)