Recently Meta has gone viral that it is working on an AI video-making algorithm named Make a Video. AI will like in picture making AI take text input and produce a video from it.
Now, a new player enters the AI video battlefield, its name is GOOGLE!
The Google research project Imagen Video is also text to video generator similar to Meta's Make a Video with one little difference. unlike Meta's offering which includes only sign-up, Google is withholding access for now under the veil of safeguarding issues.
Google has been in the Ai space before with their Imagen, a DALL-E style generator capable of photorealistic picture generation based on text prompts. The new Google tool works in a similar way but with better functions. Ai can generate a diverse range of high-fidelity videos across different aesthetic styles and with a deep understanding of 3D models thus creating high-definition, 24-frames-per-second videos.
"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,”
Jonathan Ho, author of the research paper
Despite filtering for “undesirable content”, the team admits “there is a risk that Imagen has encoded harmful stereotypes and representations”. So, until the AI - and its users - can be trusted, Google refuses to release Imagen Video without additional safeguards in place.