Create, edit and star in videos with two Google Vids updates

Gemini Omni in Vids generates and edits clips from text and image references, including stepwise changes such as background swaps and lighting fixes. Personal avatars are created from a selfie and a voice recording and then speak typed lines.

The features are for Google AI Pro and Ultra subscribers and Workspace business customers. Avatars are limited to certain regions and users 18 or older, bound to the account holder’s likeness, and generated clips carry a SynthID watermark.

This is Google’s own Workspace blog, so quality, eligibility and watermark claims are vendor-reported. The product change is real: editing becomes conversational, and presence can be synthesized from a still image and a voice sample.

Signal Desk's read: the avatar is the sharper update. Promptable editing is an expected Omni surface; an account-bound likeness in work video raises identity, consent and workplace-policy questions that a watermark does not settle.

The regional and age limits suggest Google knows the risk. Enterprise administrators will still need rules for when a synthetic stand-in is acceptable.

Vids already offered Veo 3.1 generation. Omni accepts natural-language prompts plus photos or sketches. Avatars are restricted to the account holder.