A new lawsuit is plaguing NVIDIA's rise in the tech world, with an AI copyright case from book authors claiming that its NeMo platform illegally used data for its training. Similar to OpenAI's ChatGPT copyright case, the authors claim that NVIDIA used their books for training their AI platform without these authors' authorization and consent for the large language model (LLM).
The chipmakers are now urged to pay for the damages this copyright issue caused the authors and to delete their dataset of illegally obtained books for AI training.
NVIDIA Face AI Copyright Lawsuit from Book Authors
(Photo : Noah Berger/Getty Images for Amazon Web Services)
According to a proposed class action lawsuit, NVIDIA allegedly copied the works of three authors without their consent and authorization, violating the copyright laws for the training of its AI large language model, NeMo.
The plaintiffs include Abdi Nazemian, Stewart O'Nan, and Brian Keene who argued that NVIDIA should pay for the damages it caused their names and works, as well as destroy all the copies of its Books3 dataset used for the NeMo LLM.
The authors claimed that the Books3 data set copied "all of Bibliotek," claiming that this was a "shadow library" of around 196,640 pirated books. According to Ars Technica, this was initially shared amongst the AI community, Hugging Face, but the platform defended itself stating that it is no longer accessible due to a copyright lawsuit.
Moreover, it was said that NVIDIA took the dataset from Hugging Face before it was taken down, with the company making "multiple copies" of it.
The authors are asking for a jury trial, with the court finding NVIDIA's case to have no defense for the alleged copyright infringement of their works. Moreover, they claimed that NVIDIA continued to make these illegal copies and "cause further infringement" by distributing NeMo models for their other developments in AI.
OpenAI's Copyright Lawsuit and NVIDIA
Back in July 2023, the first copyright lawsuit from book authors served up OpenAI and its ChatGPT, with authors Mona Awad and Paul Tremblay claiming that the company's large language model illegally trained on their works. Soon after, the company was hit with subsequent lawsuits from the likes of Michael Chabon, Rachel Louise Snyder, Ayelet Waldman, and David Henry Hwang.
That being said, these were taking place as NVIDIA was rising up the ranks in the realm of AI, from their renowned developments of the AI chips used for deploying the technology for the likes of OpenAI, X.AI, Microsoft, Meta, and more. With this newfound focus on AI, the company's stocks saw a 59 percent surge in artificial intelligence, skyrocketing to a $2 trillion valuation.
When NVIDIA released its AI, the company was met with enthusiasm and massive regard for its technology, but it was not long before the company faced a massive claim against its LLM training tactics. In this new AI copyright class action lawsuit, three authors came together to fight against their alleged illegal use of said works for LLM training, taking this issue to court.
