OpenAI Unveils New AI Called Sora That Creates Realistic Videos from Text
-
OpenAI unveiled Sora, a new AI that generates realistic videos from text prompts. The quality is higher than previous text-to-video models like Google's Lumiere.
-
Sora combines transformer and diffusion models to achieve coherence between video frames over up to 60 seconds.
-
Possible applications include prototyping, entertainment, advertising, and education. Future capabilities could enable world simulation.
-
Risks and concerns exist around misinformation, fake evidence, and threats from deepfakes generated by the technology.
-
Questions remain around data sourcing and intellectual property with such generative AI tools. OpenAI aims to work with experts on safety before public release.