Researchers define what counts as a world model and text-to-video generators do not
The Decoder
•
Generative AI
AI Research
An international research team wants to bring order to the fragmented world model research landscape with OpenWorldLib. Text-to-video models like Sora are explicitly left out of their definition.