Google is making some big moves in the AI arena and the latest is the ability to create videos using simple commands. The new AI model for this feature is called Lumiere which uses a diffusion model called Space-Time-U-Net which looks at the intricate details in a video and makes the changes.
Google’s AI model can actually create a clip of a few seconds using prompts which is similar to other image generating AI models we have come across in recent months.
The interesting elements that make Lumiere different from other platforms is that you can generate the content and edit as well. The magic about Google’s AI tech is that the content feels real-life like wherein you can feel a subject moving inside the water just as it would in reality. The biggest game changer for Google is the STUNet which is focused on the movement based on the timing.
Lumiere is far from reaching the public as the tests are still being done internally to make sure the AI doesn’t ruin a promising product. What also helps Google’s case with the AI model is that you can also use it for image-to-video generation where you can animate a part of the content or give it a stylised touch. Basically, the world is your canvas, in this case, thanks to Lumiere.
The biggest reason why Google is not bringing Lumiere to everyone is because the company is concerned about its misuse to generate fake content that has become a global issue, highlighted by Microsoft’s Satya Nadella and others. Google is aware of its duties to preserve the sanctity of content generated by these AI models but only time will tell how it plans to tackle the menace of fake content on the internet.
Comments
0 comment