Unveiling the Mystery: OpenAI CTO’s Startling Admission on Sora’s Training Data Origin in Viral Video
OpenAI’s CTO, Mira Murati, was grilled by The Wall Street Journal’s Joanna Stern about the training data for their new video-generating AI, Sora. Murati could only confirm that they used “publicly available data and licensed data,” dodging specific questions about sources like YouTube, Instagram, and Facebook. The lack of transparency from OpenAI’s leadership on where exactly Sora’s training data originated raises concerns about data sourcing and privacy in the age of AI development.