Google Updates Vids with Gemini Omni and Personal Digital Avatars
Google is rolling out Gemini Omni and personal avatars to Google Vids. Users can generate, edit, and swap video backgrounds with natural language prompts, and create custom talking digital avatars from a selfie.
Impact: Low
Why it matters
Developers and product leads can automate video asset generation, product updates, and demo walk-throughs using AI avatars.
TL;DR
- 01Gemini Omni enables step-by-step video editing using conversational instructions.
- 02Personal avatars can narrate text dynamically using your voice and facial likeness.
- 03SynthID watermarks are applied automatically to preserve generative transparency.
Key facts
- Supported Models
- Gemini Omni, Veo 3.1
- Required Tiers
- Google AI Pro, Ultra, Workspace Business
- Provenance Tracking
- SynthID digital watermarking
- Input requirements for avatar
- 1 selfie + 1 short voice recording
Gemini Omni and Natural Language Video Editing
Google Vids now includes Gemini Omni, allowing users to build and refine video assets step-by-step. Developers can provide a text prompt alongside reference images (such as rough UI sketches or product photos) to generate high-fidelity video tracks. Instead of restarting rendering pipelines when a change is needed, users can run incremental edit commands directly in natural language to swap environments, adjust lighting levels, or insert graphic overlays.
Secure Personal Avatars
The update also brings personal talking avatars. Subscribers of Google AI Pro, Ultra, or Workspace business plans can upload a single portrait photo and a short audio sample of their voice to generate a matching digital avatar. This avatar can then read typed scripts, eliminating the need to set up cameras or record audio for recurring developer updates, product release notes, or visual documentation.
Content Provenance and Security
To address security and safety concerns, personal avatars are strictly tied to the verified account holder's identity and likeness. Furthermore, every generated video clip is automatically embedded with Google's invisible, robust SynthID watermark, allowing downstream platforms and users to verify that the video was generated via AI.
✓ When to use
- Generating quick product release videos or internal team updates without camera setups.
- Performing quick, iterative edits on video clips using natural language commands.
✕ When NOT to use
- High-end cinematics or precise 3D rendering production that requires frame-by-frame control in professional suite tools.
What to do today
- Check your Google Workspace or AI subscription level for Google Vids access.
- Generate a test digital avatar by uploading a photo and a brief voice sample.
Sources