Google Research is working on an AI text to video generator.
The tech giant’s research branch has announced plans for Imagen Video, which would take user prompts to build looping high-definition video clips in a similar way to Meta’s recently-unveiled Make-A-Video system.
While Meta’s new offering has a sign-up option, Google is waiting until it has ironed out some safeguarding issues.
In a tweet, Jonathan Ho, an author of the research paper, said: “With the help of progressive distillation, Imagen Video can generate high-quality videos using just eight diffusion steps per sub-model.
“This speeds up video generation time substantially, by a factor of ~18x.”
The team has admitted there is a “risk that Imagen has encoded harmful stereotypes and representations”, despite trying to filter “undesirably content”.
This means Google won’t be releasing Imagen Video to the public until the team knows the system won’t be abused by the AI or its own users.
Meta recently said it was being “thoughtful” when it came to building “new generative AI systems like this”.
It said: “Make-A-Video uses publicly available datasets, which adds an extra level of transparency to the research.
“We are openly sharing this generative AI research and results with the community for their feedback, and will continue to use our responsible AI framework to refine and evolve our approach to this emerging technology.”
BANG ShowBiz Tech