Google Vids Adds Gemini Omni and Personal Avatars for Prompt-Based Video Creation

Lead

Google is turning Vids into a more conversational video creation tool. According to an update on the Google AI Blog, Google Vids is adding two features: Gemini Omni for video generation and editing, and personal avatars for creating a digital version of the user that can deliver scripted messages.

Key points

• Generate video from prompts and references: Users can describe the clip they want in natural language and add image references such as a photo or rough sketch to guide the result.
• Edit by chatting: Gemini Omni is designed not only for first drafts but also for iterative edits. Users can ask Vids to swap a background, adjust lighting or add effects to both AI-generated clips and footage captured on a phone.
• Create a personal presenter: With personal avatars, users upload a selfie and a short voice recording. They can then type a message and have the avatar deliver it without setting up a camera or recording a new take.
• Availability is limited: Gemini Omni and personal avatars are available to Google AI Pro and Ultra subscribers, as well as Google Workspace business customers. Personal avatars are limited to eligible regions and users aged 18 or older.
• AI transparency remains part of the workflow: Google says every generated clip includes an invisible SynthID digital watermark so that AI-created content can be identified.

Why it matters

The update is less about a single “text-to-video” feature and more about reshaping the video workflow. Traditional video production often requires scripting, recording, lighting, editing and narration. Google Vids is trying to compress parts of that process into a sequence of prompts, references and conversational refinements.

That could be especially useful for workplace videos: internal updates, training clips, product explainers or personalized messages. The personal avatar feature also reduces the need to appear on camera, which may help teams create video messages more quickly and consistently.

At the same time, digital likeness tools raise obvious questions around consent, identity and misuse. Google’s approach, as described in the announcement, limits personal avatars to the account holder’s likeness and adds SynthID watermarking to generated clips. Those guardrails do not eliminate every concern, but they show that discoverability and attribution are becoming central to AI video products.

Overall, Google Vids is moving from a creation surface into a broader AI-assisted production workflow: generate a draft, refine it through conversation, then deliver the message with either a real or digital presenter. The practical impact will depend on output quality, editability and how comfortable organizations are with clearly labeled AI-generated video.

Source: Google AI Blog