Thousands of documentaries are fueling AI models built by Apple, Meta, and Nvidia
Briefly

from Nieman Lab3 weeks ago
The Atlantic's investigation revealed that over 138,000 films and TV episodes were covertly used to train AI models without consent from creators, shocking the entertainment industry.
Nieman Labhttps://www.niemanlab.org/2025/01/thousands-of-documentaries-are-fueling-ai-models-built-by-apple-meta-and-nvidia/
Studio executives and directors expressed outrage, dubbing the actions of AI companies 'organized crime' for their unauthorized exploitation of Oscar-winning works and iconic productions.
Nieman Labhttps://www.niemanlab.org/2025/01/thousands-of-documentaries-are-fueling-ai-models-built-by-apple-meta-and-nvidia/
The OpenSubtitles dataset not only encompasses popular films but also includes influential documentaries by renowned filmmakers and critical news programs, further complicating copyright issues related to AI training.
Nieman Labhttps://www.niemanlab.org/2025/01/thousands-of-documentaries-are-fueling-ai-models-built-by-apple-meta-and-nvidia/
In addition to Hollywood fare, the AI model training datasets feature hundreds of documentaries and news programs, underscoring the expansive reach of AI data sourcing.
Nieman Labhttps://www.niemanlab.org/2025/01/thousands-of-documentaries-are-fueling-ai-models-built-by-apple-meta-and-nvidia/
Read at Nieman Lab
#ai-training #copyright-issues #film-industry #opensubtitles #documentary-films
[
|
]