Nvidia, OpenAI Face Legal Action from YouTube Creators

A federal lawsuit has been filed against Nvidia, a prominent player in the AI industry, alleging that the company illegally used YouTube creator David Millette’s videos to train its AI software, Cosmos. The lawsuit, which accuses Nvidia of "unjust enrichment and unfair competition," seeks to gain class action status to represent other YouTube content creators who may have been similarly affected.

According to the lawsuit, Nvidia allegedly "scraped" YouTube videos, bypassing the platform’s security measures using software on commercial servers to download vast amounts of content—around 80 years’ worth of video content per day. These claims are supported by an August 5 report from 404 Media, which was cited in the legal filing. The lawsuit was submitted in the Northern District of California on Wednesday.

In response to the allegations, Nvidia stated, "We respect the rights of all content creators and are confident that we are working in full compliance with the letter and spirit of the law." This defense echoes Nvidia’s position that its actions were lawful and aligned with the principles of creating transformative works from publicly available information.

The lawsuit against Nvidia follows closely on the heels of a similar legal action Millette filed against OpenAI. These lawsuits underscore the massive data requirements for training large language models (LLMs) that power Generative AI (GenAI), with companies often sourcing data from publicly accessible online content. However, Millette’s lawsuits do not accuse the tech companies of copyright infringement; instead, they focus on claims of unjust enrichment and unfair business practices. In Nvidia’s case, the lawsuit argues that the company's methods were “unfair, immoral, unethical, oppressive, unscrupulous, or injurious to consumers.”

Millette is seeking an injunction to prevent the unauthorized use of his videos, along with restitution and other damages. The complaint also alleges that Nvidia obtained video content not just from YouTube but also from other platforms like Netflix, as reported by 404 Media. This report includes claims from a former Nvidia employee, who provided information under the condition of anonymity. Nvidia, in its defense, asserts that learning from publicly available facts and ideas is a legitimate and encouraged practice within the legal framework. The company’s stance emphasizes that creating new and transformative works based on such learning is both fair and just.

The broader context of this legal action reflects the increasing scrutiny of how AI companies source and use data for training their models. For instance, an April report by The New York Times revealed that OpenAI had developed a speech-recognition tool called Whisper, which was used to transcribe audio from YouTube videos to train its language models after exhausting other online text sources. This highlights the significant role that publicly available multimedia content plays in advancing AI technologies.

The legal challenges come at a time when Nvidia’s stock has seen a dramatic surge—up by approximately 600% since early 2023—driven by its association with the AI boom. Nvidia's market capitalization has now exceeded $3 trillion, solidifying its position as a key player in Silicon Valley’s tech landscape. However, these lawsuits could introduce new challenges as Nvidia navigates the legal and ethical complexities of AI development.