OpenAI copyright case draws 51 briefs as Brave is sued and EU splits training from answers

A dozen cases, 51 briefs and a split bar

The split is lopsided but not decisive: 26 briefs support the plaintiffs, 18 support the defendants and eight are neutral.

Amicus briefs let parties who are not in a case ask the court to hear their view. What Judge Sidney H. Stein decides will shape what publishers can charge AI developers for news and books, and what AI firms must pay to keep training on scraped material. The exposure reaches buyers of AI tools too: Adobe says it will cover customers' legal costs if content made with its Firefly models is challenged, subject to contract terms.

The case consolidates more than a dozen suits, including actions filed by the Authors Guild and the New York Times. Both sides filed cross-motions for summary judgment last month. The plaintiffs argue OpenAI infringed by copying works to train its models, including from pirated sources. They also argue its outputs can infringe. OpenAI counters that copying for training is fair use and that its outputs do not infringe. The US Department of Justice filed a Statement of Interest on 4 September urging the court to find training on copyrighted works fair use. In October Stein struck from the evidence a study suggesting AI books were displacing human-written works. OpenAI had argued the study's author failed to disclose a funding conflict; the author denies it, and the plaintiffs did not oppose the motion.

Brave accused of selling a back door to news

The Chicago Tribune, the New York Times and Tribune Publishing papers the Daily News and the Orlando Sentinel filed federal suits in Manhattan on Monday against Brave Software.

The complaint calls Brave "the supplier, dealer, and distributor of stolen copyrighted content", resold to AI firms including Anthropic and Perplexity. It says Brave crawled and copied content from more than 30 billion webpages without identifying its crawler or giving site owners a way to block it. It lists five claims, covering direct, contributory and vicarious copyright infringement. In one example, a user asked Brave's "Ask Brave" feature for articles published that day in the Daily News. The chatbot listed the titles of 35 paywalled articles and offered to fetch the full text of whichever one the user picked.

Frank Pine, executive editor at Tribune Publishing and MediaNews Group, said the software company had made "a mockery of our intellectual property laws".

Steven Lieberman, an attorney for the Daily News and its sister papers, said: "Buying stolen goods secondhand doesn't make them legitimate".

Europe splits training from answering

SecNews reported on 3 November on an opinion from Advocate General Maciej Szpunar in case C-250/25, the CJEU's first case on generative AI and copyright.

Szpunar treats training and answering as separate acts. Using protected works to train large language models may be permitted under the text and data mining exception in European copyright law, the opinion says. Reproducing protected expression when a chatbot answers a user may infringe regardless of whether the training was lawful. He drew a further line: facts in a news article are not protected. The author's original expression, meaning how the material is organized and presented, is. The case was brought by Hungarian publisher Like Company against Google Ireland, which claims Gemini produced summaries of its articles without permission.

What happens next

A final round of briefs in the OpenAI case is due on 6 November. Oral arguments will then be set before Stein.