NVIDIA in tough trouble: they sue it for infringing copyright in its AI

0
22
nvidia in tough trouble they sue it for infringing copyright.jpg
nvidia in tough trouble they sue it for infringing copyright.jpg

NVIDIA NeMo

The software company NVIDIA, known for its graphics processing units (GPUs), has a platform focused on artificial intelligence (AI) and data science called ‘NeMo’. According to some users, this service allows companies to create and train their own chatbots with content copyrighted and, for this reason, they have recently been sued.

The writers Abdi Nazemian, Brian Keene and Stewart O’Nan, among others, have filed a class action lawsuit against NVIDIA, as leaked by the media outlet Ars Technica. The document states that the company shared data from its books illegally, since it did not have its consent.

Almost 200,000 pirated books in AI

The lawsuit claims that the large language model (LLM) ‘NeMo’ has been trained with illegal content from Books3a pirate library with a catalog of some 196,640 books. This AI could also be used “as a basis from which to build more models”, so there could be other chatbots that violate copyright.

The Books3 data set was once shared from the Hugging Face website, but “is obsolete” since October 2023, since it violates copyright. Despite this, the plaintiffs assure that NVIDIA made “multiple copies” of that information to train your chatbot.

Books3 copyright

To defend their claim, the novelists refer to the fact that the firm “has admitted to having trained its NeMo Megatron models in a copy of The Pile data set». They emphasize that Books3 is part of The Pile, so they trained their AI with “one or more copies of the infringed works” that were within the catalog of the pirate data set.

Additionally, the writers claim that since ‘NeMo’ can be used “as a base from which to build more models,” other chatbots may be found to be in breach of copyright.

An NVIDIA spokesperson has defended the company in an article in The Wall Street Journal, claiming that they respect “the rights of all content creators and believe that we created NeMo in constant compliance with copyright law.”

With the class action lawsuit, the affected authors hope that it will be carried out a jury trial and it will be a court that verifies whether or not NVIDIA fails to comply with the regulations.

OpenAI was also sued for the same

NVIDIA’s case is not unique. On other occasions, there have already been problems with copyright and the training of tools with artificial intelligence. Possibly, one of the class actions that has generated the most stir is that of OpenAI from last year.

The situation at Sam Altman’s company was similar. The authors stated that infringed copyright by training ChatGPT with its books and allowing it to reference them in its responses.

Artificial intelligence trained with books

Additionally, the plaintiffs claimed that OpenAI’s AI had a “strange ability to generate similar text to that found in textual materials protected by copyright. In this case, they showed some concern about possible attempts at plagiarism.

Recently, Araceli Martínez-Olguín, a California district judge, dismissed many of the authors’ allegations for lack of supporting evidence. The plaintiffs had until today, March 13, to modify their arguments and continue with their claims.

Previous articleThis WhatsApp novelty will help you distinguish individual chats from groups
Next articleWhatsApp blocks screenshots so you can’t ‘steal’ profile photos
Abraham
Expert tech and gaming writer, blending computer science expertise