Gemini lets friends share avatars for AI images and videos

Google is rolling out shareable personal avatars in the Gemini app. A user can create a reusable digital version of themselves, then use it in generated images and videos without uploading a new reference each time.

The new sharing layer lets friends grant one another access and tag approved avatars directly in prompts. Google says Gemini maintains character and voice consistency, making it easier to place the same people across multiple scenes instead of rebuilding their likeness for every generation.

Control stays with the avatar owner: participation is optional, and access can be added or removed. Google’s Gemini Omni page says personal avatars are designed to produce video that looks and sounds like the user. It also says videos generated in Gemini include the company’s invisible SynthID watermark, while Gemini can inspect uploaded media for SynthID and other signs of AI generation.

The feature brings a “cameo” workflow into Gemini’s main creative tools. Its importance is not just convenience. Reusable likenesses and cloned voices make consent and revocation part of the product itself, rather than something users must negotiate outside the app.

Google has not detailed availability by country or subscription tier in the announcement. Its product page says access to Gemini Omni features varies by plan and region, so the sharing option may not appear for every user immediately.