NVIDIA allegedly scraped Netflix and YouTube videos without permission to train its AI

Updated on 06-Aug-2024
HIGHLIGHTS

NVIDIA is reportedly building an AI video foundational model that is said to have scraped copyrighted video content.

While working on this model, many NVIDIA employees asked their managers about legal and ethical issues.

According to a 404 Media report, NVIDIA has scraped a lot of copyrighted video content from YouTube and Netflix to train its AI model.

NVIDIA has found itself in the middle of a controversy. The tech giant is reportedly building an AI video foundational model that is said to have scraped copyrighted video content from YouTube and Netflix. This information spread like wildfire when conversations between NVIDIA employees on Slack were leaked by 404 Media. While working on this model, many NVIDIA employees asked their managers about legal and ethical issues, and here’s what happened

404 Media did an exclusive investigation and spoke to an anonymous former NVIDIA employee who revealed internal NVIDIA processes. According to that, NVIDIA has scraped a lot of copyrighted video content from YouTube and Netflix to train Cosmos, the AI video foundational model. NVIDIA is building this AI model to combine light simulation, physics, and intelligence in one system to enable many important applications for NVIDIA.

The leaked conversation shared by Samantha Cole, the author who wrote this copy on 404 Media, reveals NVIDIA employees raising concerns over ethical and legal issues. To this, Ming-Yu Liu, Vice President of Research at NVIDIA soon replied, “This is an executive decision. We have an umbrella approval for all of the data.” 

Apart from that, many more employees reportedly also asked the same questions about legality in the process of training Cosmos. However, their project managers told them that they have the authority approval to do such work. Which authority? We don’t know that yet. 

Additionally, the report also stated that NVIDIA’s Cosmos team talked about downloading different types of video content, such as Hollywood movies, Discovery Channel documentaries, and high-quality gaming footage, in a Slack conversation. Project leaders like Ming-Yu Liu said that Hollywood films could be useful for training because of their “gaming-like 3D consistency and fictional content but much higher quality.” However, he also noted the sensitivity of using this content, similar to the concerns raised by artists after the release of Stable Diffusion (SD). 

According to Cole, NVIDIA is consuming copyrighted content as fast as it can to “stay competitive in the AI industry.” If this is the case, this is not going to be a fair competition then. We haven’t heard anything from NVIDIA about this yet, so let’s see how it is planning to counter these allegations.

Mustafa Khan

Mustafa is new on the block and is a tech geek who is currently working with Digit as a News Writer. He tests the new gadgets that come on board and writes for the news desk. He has found his way with words and you can count on him when in need of tech advice. No judgement. He is based out of Delhi, he’s your person for good photos, good food recommendations, and to know about anything GenZ.

Connect On :