Google Adds Avatar Tool and New Video Model to Flow AI App

Google updated its Flow AI video creation tool with a new video-generation model and a feature that lets users create digital versions of themselves for insertion into AI-generated clips. The company announced the changes at its annual I/O developer conference in Mountain View, California.

Elias Roman, vice president of product management at Google Labs, demonstrated the new “avatar” feature during the event.

The update replaces the previous Veo model with Omni Flash.

Roman said the new model improves consistency when generating characters across multiple video clips. Users can now generate scenes with more detailed visuals and maintain character appearance through successive generations, according to the demonstration.

To create an avatar, users scan a QR code in their Flow account settings and record themselves saying a sequence of numbers while moving their head to capture different angles. The resulting digital version can be placed into AI-generated videos. Every clip created with the Omni Flash model includes Google’s SynthID watermark.

Roman said the tool is intended for creators who want to appear in content without filming themselves. The avatars are also available through the Gemini app and YouTube. Google launched Flow last year under its experimental Labs division. Last month, YouTube Shorts introduced a similar limited option for AI avatars in short-form clips.