Tech giant Nvidia is facing a lawsuit from a group of authors who accuse it of using their copyrighted works without permission to train its AI platform NeMo.
Brian Keene, Abdi Nazemian and Stewart O’Nan said their works were part of a dataset of about 196,640 books that had helped train NeMo to simulate ordinary written language, and were taken down in October “due to reported copyright infringement.”
The authors’ proposed class action lawsuit was filed Friday in San Francisco federal court and claims that Nvidia has “admitted” it trained NeMo on the dataset, thus infringing their copyrights. It seeks unspecified damages for people in the U.S whose copyrighted works have helped train NeMo’s large language models (LLMs) in the past three years.
Among the many works included in the lawsuit are Keene’s 2008 novel “Ghost Walk,” Nazemian’s 2019 novel “Like a Love Story,” and O’Nan’s 2007 novella “Last Night at the Lobster.”
The suit launched by the three authors claims the books were included in a data known as “The Pile” that contained a collection of books called “Books3”, and that Nvidia has previously admitted to training its NeMo Megatron AI models on The Pile and Books3.
This lawsuit joins the ranks of a growing group of lawsuits against tech companies ver the use of copyrighted content in training AI models, including several filed by writers as well as by the New York Times, which sued ChatGPT-maker OpenAI and Microsoft.
Nvidia declined to comment on the lawsuit against them.
Leave feedback about this