Cohere faces lawsuit from major publishers over copyright infringement

Major publishers, including Politico, Vox, The Atlantic, and The Guardian, have filed a lawsuit against the AI startup Cohere, claiming copyright and trademark infringement. This legal battle underscores the ongoing tensions between original content creators and AI technologies that utilize their work without permission.

According to the claim reported by the Wall Street Journal, the lawsuit alleges that Cohere improperly incorporated over 4,000 copyrighted works into its training data for a large language model. Additionally, the complaint states that recipients of Cohere’s AI-generated outputs were not given proper attribution, leading to misunderstandings about the authorship of the content.

Danielle Coffey, CEO of the News Media Alliance, criticized Cohere, asserting, “Rather than create their own content, they’re stealing ours to compete with us without our permission, without compensation.” This statement highlights a central concern of the publishers: the encroachment by AI on their business models.

The suit also mentions instances of trademark infringement, where Cohere’s AI is purported to provide attributes to articles, yet the content is riddled with inaccuracies. A notable incident cited involved an AI-generated piece that erroneously mixed details of a terror attack with a separate shooting, creating disinformation that potentially damages the reputations of both the AI firm and the news organizations it mimics.

The publishers are seeking substantial financial compensation under the Copyright Act, aiming for damages of $150,000 for each infringed work. Furthermore, they aspire to curtail Cohere’s access to proprietary content and establish clearer terms regarding the use of journalism in AI training processes. As emphasized by Pam Wasserstein of Vox Media, this lawsuit could set critical legal precedents for the relationship between AI and journalism.

Cohere responded to these allegations, stating it stands firmly behind its practices to responsibly train its AI technology, asserting that the lawsuit is both