NVIDIA’s AI team reportedly scraped YouTube, Netflix videos without permission

NVIDIA reportedly scraped YouTube, Netflix videos for AI training without permission, raising ethical and legal concerns.

: NVIDIA allegedly scraped videos from YouTube, Netflix, and other sources to train AI models without permission. The practice raised significant ethical and legal issues, especially regarding IP laws. YouTube spokesman Jack Malon emphasized that scraping violates YouTube's terms. NVIDIA staff reported the practice was approved by top executives despite concerns.

NVIDIA is reported to have used videos from YouTube, Netflix, and other datasets for AI training without obtaining permissions. The company aimed to develop commercial AI projects, including self-driving cars, the Omniverse 3D world generator, and 'digital human' efforts. This practice stirred ethical and legal issues, although NVIDIA defended itself by claiming its actions were compliant with copyright laws, stating that IP laws protect expressions, not facts or data.

YouTube disagreed with NVIDIA’s stance, with spokesperson Jack Malon referencing past statements from YouTube’s CEO Neal Mohan. Mohan had asserted that using YouTube content for AI training constitutes a clear violation of terms. Concerns were echoed when reports emerged that companies such as OpenAI and Runway AI were similarly implicated.

NVIDIA employees raising legal and ethical doubts were reportedly told the scraping had high-level executive approval. The company allegedly downloaded content using virtual machines and rotating IP addresses to avoid bans. Despite internal concerns, NVIDIA proceeded, believing they had 'umbrella approval' for their data use methods, including those datasets initially meant for academic use.

AI TrainingVideo scrapingNvidia