The battle between AI start-ups on one side and more traditional media outlets on the other shows no signs of abating. 

A new front was opened last week when a dozen leading news and magazine publishers including Condé Nast, Forbes, The Atlantic, the LA Times and the Guardian, filed a lawsuit at the US District Court for the Southern District of New York against AI start-up Cohere alleging copyright and trademark infringement.

News of the suit came hard on the heels of a notable victory by Thomson Reuters in its closely watched copyright battle with AI legal research start-up ROSS (see commentary below).

The latest suit – described as “misguided and frivolous” by Cohere – alleges the start-up “engages in wholesale, verbatim copying” of the publishers’ works to train its models. 

The plaintiffs allege Cohere uses scraped copies of articles, through training, real-time use and in outputs “without permission or compensation” to power its AI service, which in turn “competes with publisher offerings and the emerging market for AI licensing”.

They also accuse Cohere of manufacturing “fake pieces” which are attributed to them, “misleading the public and tarnishing our brands”.

Founded in 2019, Cohere develops, operates and licences AI models. According to court documents, as of July 2024, Cohere was valued at $5.5bn. Its lead funders include other large corporate technology companies including Oracle, NVIDIA and Salesforce.

All the plaintiffs are members of leading industry trade organisation News/Media Alliance and are represented in the case by litigation boutique Oppenheim and Zebrak’s Scott Zebrak, Jenny Pariser and Meredith Stewart, facilitated by Regan Smith, general counsel of the News/Media Alliance.

Danielle Coffey, president and CEO of the News/Media Alliance said: “As news, magazine and media publishers, we serve an important role in keeping society informed and supporting the free flow of information and ideas, but we cannot continue to do so if AI companies like Cohere are able to undercut our businesses while using our own content to compete with us.”

Anna Bateson, CEO of Guardian Media Group, added that as part of a considered approach to generative AI “the Guardian has explored and signed agreements with numerous partners to ensure fair compensation and attribution for the Guardian’s award-winning investigative journalism”.

She added that unfortunately, Cohere has demonstrated an “egregious pattern of scraping and copying news articles to produce full verbatim copies of original content without compensation – or even worse, complete hallucinations. The Guardian is proud to stand with some of the world’s top publishers in an attempt to stop Cohere’s brazen theft and distortion of original journalism”.

A Cohere spokesperson told the Canadian website BetaKit that it “strongly stands by its practices for responsibly training its enterprise AI”.

“We have long prioritised controls that mitigate the risk of IP infringement and respect the rights of holders,” the spokesperson said.

“We would have welcomed a conversation about their specific concerns – and the opportunity to explain our enterprise-focused approach – rather than learning about them in a filing. We believe this lawsuit is misguided and frivolous, and expect this matter to be resolved in our favour.”

The publishers are seeking injunctive relief and damages under the US Copyright Act and US Lanham (trademark) Act, as well as a trial by jury.

Slew of lawsuits

The lawsuit was filed two days after judgment was handed down in the first major US AI copyright case when Thomson Reuters prevailed over AI legal research start-up ROSS over its use of copyrighted Westlaw headnotes and other materials as training data for its search engine. 

Delaware district court judge Stephanos Bibas reversed his summary judgment opinion of 2023, granting Thomson Reuters’ motion for summary judgment on direct copyright infringement and related defences, as well as fair use. He rejected ROSS’s arguments that its use of works was protected by the fair use doctrine.

Commenting on the case, Debevoise & Plimpton partner Megan Bannigan, counsel Christopher Ford and associates Samuel J Allaman and Abigail Liles pointed out that the AI technology utilised by ROSS was search-based as opposed to generative AI, which creates new content based on user prompts. But they said it nevertheless “provides a helpful early barometer for how future courts will evaluate copyright infringement claims, and the fair use defence, in the context of AI technologies”.

On Tuesday (18 February), there was an update on AI search engine Perplexity AI’s ongoing legal wrangle with Wall Street Journal and New York Post publishers Dow Jones & Co and NYP Holdings. Bloomberg reported that the start-up urged a federal judge to throw out the copyright infringement suit arguing that the Manhattan court lacks jurisdiction over the California-based search engine.