Google Vids introduces the Gemini Omni model. You can customize your own digital avatar by uploading a selfie + audio. Even though OpenAI’s Sora may have retired, Google still believes that users’ obsession with appearing in AI videos has not cooled down.

Google Vids introduces the Gemini Omni model, and you can customize your own digital avatar by uploading a selfie + audio. Even though OpenAI’s Sora may have retired, Google still believes that users’ obsession with appearing in AI videos has not cooled down. On Thursday, the technology giant launched a blockbuster update for Google Vids: You only need to upload a selfie and a voice recording, and the system will create an exclusive digital virtual person that closely matches you in appearance and voice. In other words, there is no need for a real person to be present in front of the camera, a digital clone can speak for you. What debuted simultaneously with the digital human was Google’s own multi-modal model Gemini Omni, which was integrated into the Vids platform. This combination turns video generation into a dialogue between text and pictures - users can produce their own videos by combining text prompts with uploaded reference pictures. Omni has also packed in several practical skills: changing the background, modifying light, and adding special effects to videos taken casually with a mobile phone are all a breeze. What's even more pleasing is that it now supports progressive editing. You can make changes step by step during creation. If you are not satisfied, you don't have to start over from scratch, saving you the torture of starting from scratch repeatedly. This series of upgrades is dragging Google Vids from an AI-assisted office demonstration tool to an all-round video creation platform. By incorporating Vids into Google Workspace, the signal released by Google is very straightforward: it wants to become a business tool for companies to make internal announcements, training videos and other scenarios. However, the two new cards of personalized virtual humans and conversational editing will inevitably put it on the same track as AI video startups such as HeyGen, Synthesia, Captions, and D-ID. When it comes to security and compliance, Google is thinking ahead. It particularly emphasized that the new AI virtual person will be strictly bound to the account holder's real appearance and his Google account, and will be marked with an invisible watermark through SynthID technology. This line of defense means you can’t use it to create spoofs or fake videos of others, such as a face-swapping prank on Google CEO Sundar Pichai. Currently, access to the personal avatar function is only open to users in specific regions who are over 18 years old. When AI clones begin to enter office suites, what Google is trying to prove is that convenience and boundaries can coexist. via AI News (author: AI Base)