A group of publishers are suing AI company Cohere, claiming the GenAI startup carried out major copyright infringements.
The publishers, including The Guardian, Vox Media, Politico, Forbes, Condé Nast and The Atlantic, claim Cohere used over 4,000 copyrighted works to train its LLM and display articles for users without proper attribution, which has hindered traffic on their websites.
The lawsuit, first reported by The Wall Street Journal, also claims the AI tool infringed on publishers’ trademarks by ‘hallucinating’ content that was never published.
One example outlined in the suit references an article published in The Guardian about Hamas’s attack on the Nova music festival in Israel, where the AI model confused the event with a 2020 shooting in Nova Scotia, Croatia.
The publishers are seeking $150,000 per copyrighted work, the maximum amount for damages. The lawsuit also wants to reduce the access Cohere has to articles and establish terms for how AI can use journalistic content for training.
Cohere has defended itself, saying it “expect[s] this matter to be resolved in our favour”, and referred to the lawsuit as “misguided and frivolous”.
“We have long prioritised controls that mitigate the risk of IP infringement and respect the rights of holders,” said Cohere in a statement. “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”.
Cohere is a Canadian AI startup with headquarters in Toronto and San Francisco. The company is currently valued at $5 billion and is known its chatbot and for building AI applications for businesses.
The Cohere suit is only one case of legal action against AI companies for alleged copyright violations. The New York Times famously sued ChatGPT owner OpenAI in 2023 for copyright infringement, and last week Thomson Reuters won a copyright lawsuit against AI startup Ross Intelligence.
