Nvidia is now facing a copyright challenge tied to AI training, as authors Brian Keene, Abdi Nazemian, and Stewart O'Nan claim their works were included in a dataset used to train the company's NeMo AI framework.
The case adds Nvidia to a growing group of major AI companies accused by authors of using books without permission to build or improve large language models.
What The Authors Claim
The authors say their works were part of a dataset of approximately 196,640 books used to train Nvidia's NeMo AI framework. The central issue is whether copyrighted books were used in AI training without the authors' permission.
The lawsuit is a class action filed Friday night in San Francisco. According to the source article, the authors are seeking damages for individuals in the U.S. whose copyrighted works were used to train NeMo's LLMs over the past three years.
That scope matters because the complaint is not presented as a dispute over one book or one author alone. It is framed around a broader group of people in the U.S. whose copyrighted works may have been included in training data for NeMo's large language models.
Why NeMo Is At The Center
Nvidia's NeMo is described as an AI framework that allows companies to customize, deploy, and fine-tune pre-trained large models. In plain terms, the framework is connected to the practical work of adapting large models for use by companies.
The authors' claim focuses on the training side of that system. Their position, as described in the source article, is that their books were part of the material used to train NeMo's LLMs.
That makes the dataset a key part of the dispute. The article identifies the dataset as including approximately 196,640 books, and the authors claim their works were among them. The question raised by the lawsuit is not simply whether Nvidia offers AI tools, but whether copyrighted books were used to train those tools without permission.
Part Of A Wider AI Copyright Battle
The Nvidia lawsuit fits into a larger pattern involving authors and AI companies. Several authors have filed lawsuits against AI companies such as Meta, OpenAI and Microsoft, accusing them of copyright infringement.
Those claims, as summarized in the source article, are built around the allegation that books were used without permission to train AI models such as GPT-4. Nvidia is now named in a similar kind of dispute, though this case centers on NeMo's LLMs.
The broader issue is straightforward: large language models require training, and authors are challenging how some companies obtained or used the text behind that training. The cases mentioned in the source article involve claims about books, permission, and copyrighted works.
- Authors named in the Nvidia case: Brian Keene, Abdi Nazemian, and Stewart O'Nan.
- Company named: Nvidia.
- AI framework at issue: Nvidia's NeMo AI framework.
- Dataset size cited: approximately 196,640 books.
- Relief sought: damages for individuals in the U.S. whose copyrighted works were used to train NeMo's LLMs over the past three years.
What Is At Stake
For authors, the core concern is control over copyrighted work. If a book is used to train an AI model, the author may argue that permission should have been required before that use took place.
For AI companies, the dispute touches the data behind large models and the process used to train them. Nvidia's NeMo allows companies to customize, deploy, and fine-tune pre-trained large models, so allegations about the training material behind NeMo's LLMs go directly to the foundation of the framework.
The source article does not report an outcome in the Nvidia case. It reports the filing, the authors' claims, the dataset size, the AI framework involved, and the damages sought for a proposed class of U.S. individuals.
That makes this lawsuit another important marker in the ongoing conflict over AI training and copyrighted books. The facts now presented are allegations from the authors, and the case places Nvidia alongside Meta, OpenAI and Microsoft as companies facing author-led copyright infringement claims connected to AI models.