Authors File Lawsuit Against Nvidia for Copyright Infringement in AI Training

Three authors, Brian Keene, Abdi Nazemian, and Stewart O’Nan, have initiated a lawsuit against Nvidia, alleging the company used their copyrighted books without permission to train its NeMo AI platform. The lawsuit, filed in San Francisco federal court, revolves around a dataset comprising approximately 196,640 books utilized to refine NeMo’s capability to mimic ordinary written language. This dataset was removed in October following copyright infringement reports, an action the authors interpret as Nvidia’s admission of copyright infringement.

The plaintiffs are pursuing unspecified damages on behalf of U.S. individuals whose copyrighted works were employed in training NeMo’s large language models over the past three years. The works specified in the lawsuit include Keene’s "Ghost Walk," Nazemian’s "Like a Love Story," and O’Nan’s "Last Night at the Lobster." Nvidia has refrained from commenting on the lawsuit, and the authors’ legal representatives have not yet provided additional remarks.

This legal action against Nvidia joins a growing series of lawsuits targeting the use of generative AI technologies that produce new content based on existing texts, images, and sounds. Companies such as OpenAI, creators of ChatGPT, and their partner Microsoft have also faced similar lawsuits. Amidst this backdrop, Nvidia’s stock has surged nearly 600% since the end of 2022, propelled by the AI boom, bringing the company’s market valuation to almost $2.2 trillion. This case highlights the burgeoning legal challenges facing the AI industry, particularly concerning copyright and intellectual property rights.