Sora, the AI model created by OpenAI, continues to raise questions about the data used to create high-quality, long videos. Many wonder what specific data, particularly large video datasets like those from YouTube, Sora utilizes for training.
Previously, Mira Murati, the CTO of OpenAI, refused to confirm whether Sora used YouTube videos for training, prompting the YouTube CEO to suggest that doing so would violate usage terms. This question resurfaced at the Bloomberg Technology Summit, where Brad Lightcap, the COO of OpenAI, was asked if Sora is trained using YouTube videos.
Lightcap responded by emphasizing OpenAI’s commitment to transparency regarding data sources, such as the Content ID system, allowing content creators to control access to their data for AI training. OpenAI is actively seeking ways to engage with creators, publishers, and stakeholders to address concerns and safeguard their interests. This ongoing issue remains unresolved, with no definitive answer at this time.
TLDR: Sora, the AI model from OpenAI, is under scrutiny for its use of data sources, particularly large video datasets like those from YouTube. The debate underscores the importance of transparency and collaboration between AI developers and content creators.
Leave a Comment