Subscribe to Teknoblog content on Google News:
It is becoming increasingly difficult to follow the cases filed in the courts against productive artificial intelligence technologies. Recently, The New York Times took Microsoft and OpenAI to court. Finally, a lawsuit filed against Nvidia by a group of authors was added to this group.
The lawsuit alleges that Nvidia used authors’ books without permission to train its AI platform NeMo, a language model that allows businesses to build and train their own chatbots.
Authors Abdi Nazemian, Brian Keene, and Stewart O’Nan requested a jury trial and asked Nvidia to pay damages and destroy all copies of the Books3 dataset used to support NeMo large language models (LLMs). The authors also claim that the data set copied the shadow library called Bibliotek, which consists of 196 thousand 640 pirated books.
The petition includes the following statements: “In summary, Nvidia admitted that it trained NeMo Megatron models on a copy of The Pile dataset. That’s why the company made sure to train its NeMo Megatron models on a copy of Books3 as well. Because Books3 is part of The Pile. Certain books written by plaintiffs are part of Books3, including the infringing works, and thus the company necessarily trained its NeMo Megatron models on one or more copies of the infringing works, thus directly infringing plaintiffs’ copyrights.”
Nvidia does not accept the claims
Nvidia’s statement to The Wall Street Journal regarding the case is as follows: “We respect the rights of all content creators and believe that we created NeMo in full compliance with copyright law.”