Google shows off Lumiere, a space-time diffusion model for realistic AI videos 

Lumiere was trained on a dataset of 30 million videos, along with their text captions, and is capable of generating 80 frames at 16 fps. The source of this data, however, remains unclear at this stage.