Google Vids Adds Personal Avatars: One Selfie + One Voice Clip Makes a Talking "You"
Google added personal digital avatars to Google Vids — upload one selfie and one voice clip to generate a look-and-sound-alike; the multimodal Gemini Omni also arrives, supporting conversational, step-by-step editing.
One selfie + one voice clip becomes a "digital avatar"
Per IThome (July 17), Google added a personal avatar feature to Google Vids: users upload one selfie and one voice recording to generate a virtual persona matching their look and voice, starring in their own "AI films."
Enter Gemini Omni: conversational, step-by-step editing
The multimodal Gemini Omni model is also coming to Google Vids. Users can submit text prompts plus reference images for Omni to compose videos; beyond building from scratch, Omni can swap phone-video backgrounds, improve lighting and add effects, and supports step-by-step edits — adjusting as you go without regenerating the whole clip. Originally a workplace-presentation tool, the updated Vids is closer to a full video-creation platform, entering the AI-video arena against HeyGen, Synthesia, Captions and D-ID.
Anti-abuse: your face only, plus SynthID watermark
Google stresses avatars can only be generated in the account holder's own likeness and are bound to that Google account, with outputs carrying an invisible SynthID watermark. TechCrunch notes this means users can't make a spoof video of Google CEO Sundar Pichai — a contrast with OpenAI's Sora period, when users could "prank" Sam Altman. For now it's limited to users 18+ in select regions. The "bind-to-self plus mandatory watermark" design reflects how big tech is balancing avatar misuse and compliance pressure.