Authors Sue Nvidia for Use of Copyrighted Works in AI

Read Time:1 Minute, 9 Second

Some authors have sued Nvidia, a company that makes chips that power AI, saying that it used their stolen works to train its NeMo AI platform without their permission.

A set of about 196,640 books that helped NeMo learn how to write generally included works by Brian Keene, Abdi Nazemian, and Stewart O’Nan. The dataset was taken down in October “due to reported copyright infringement.”

The authors said that the takedown shows that Nvidia “admitted” using the dataset to train NeMo, which was against their copyrights. They filed the proposed class action on Friday night in federal court in San Francisco.

“In the last three years, individuals in the United States whose copyrighted materials aided in the training of NeMo’s ‘large language models’ are now facing a lawsuit seeking unspecified damages.”

“Books like “Ghost Walk” authored by Keene in 2008, “Like a Love Story” written by Nazemian in 2019, and “Last Night at the Lobster” by O’Nan are some of the titles involved in the legal case.”

On Sunday, Nvidia wouldn’t say anything. On Sunday, lawyers for the authors sought further details but did not immediately respond.

There are more and more lawsuits against Nvidia from writers and the New York Times because of generative AI, which takes text, images, and sounds and makes new content.

GeForce says NeMo is an easy and quick way to use generative AI.