404 Media was able to gain access to classified NVIDIA documents that described details of the project codenamed Cosmos. Every day, the company’s employees downloaded huge amounts of video content from various online services such as YouTube and Netflix. The total amount of processed material was impressive – it was equal to 80 years of continuous viewing.
The goal of this effort was to use the collected data for several ambitious AI-powered projects at NVIDIA, including creating 3D worlds in Omniverse, improving self-driving car technology, and developing digital assistants.
After the information appeared, NVIDIA gave an official response. The company claims that they do not violate copyright law, since they use only ideas and information, not finished products. In addition, they believe that training artificial intelligence falls under the concept of “fair use”. However, the owners of the resources from which the content was downloaded do not agree with this approach.
It is not yet known whether a class action lawsuit will be filed in court.
