Google's video tool Vids gained its most personal feature yet on July 16: AI avatars of the user themselves, generated from a selfie and a short voice recording, that will deliver any script on camera — no filming, no reshoots, no camera shyness.
How it works
Users upload a selfie and record a brief voice sample; Vids builds an avatar that looks and sounds like them, ready to present whatever script it's given. The obvious use cases are the unglamorous backbone of corporate video: training modules, product walkthroughs, onboarding clips and announcement videos that previously required either an on-camera performance or a contractor.
The guardrails
Google has drawn the consent lines tightly. Avatars are restricted to the account holder's own likeness and tied to the Google Account; the feature is limited to users 18 and older, available in certain regions, and every generated clip carries an invisible SynthID watermark for provenance detection. The design goal is explicit: self-impersonation as a productivity feature, third-party impersonation ruled out at the account level.
Gemini Omni joins
Alongside avatars, Google's Gemini Omni generation model comes to Vids, bringing prompt-based creation and editing: build video from written descriptions and reference images, swap backgrounds, fix lighting on phone-recorded footage and apply effects — with support for step-by-step iterative edits that refine a video without starting over.
Who gets it
The features are rolling out to Google AI Pro and Ultra subscribers and Workspace business customers. The move squeezes dedicated avatar vendors like HeyGen and Synthesia from the platform flank: what they sell as a product, Google now bundles into the Workspace stack most of their target customers already pay for — TechCrunch's framing that Vids is becoming an all-in-one video platform rather than a presentation tool looks increasingly literal.