Google Vids Launches Personal AI Avatars, Competing with Specialized Video Tools

Google has announced a substantial upgrade to Vids, its AI-powered video creation tool, introducing the ability for users to generate customized digital avatars that replicate their likeness and vocal characteristics. The move comes as the video generation space heats up following OpenAI's decision to discontinue its Sora app.

Personal Avatars and Expanded Capabilities#

Users can now create a bespoke avatar by uploading a selfie and a voice sample, producing a digital representation that captures both visual and audio qualities. The platform has also integrated Gemini Omni, Google's multi-modal AI model, which enables video generation using a blend of text prompts and reference images. This combination approach gives creators more flexibility in specifying what they want their videos to look like.

Beyond avatar generation, the Omni integration provides additional editing powers: users can modify backgrounds, adjust lighting in existing phone-recorded footage, and layer on visual effects. The system now supports incremental edits as well, letting creators refine videos step-by-step rather than starting over entirely when adjustments are needed.

Positioning as a Comprehensive Creation Tool#

The updates represent a deliberate expansion of Vids' original purpose. Previously framed as a Workspace-focused tool for corporate presentations and internal communications, the platform is now positioning itself as a full-featured video creation suite. This broader ambition puts Google in more direct competition with established AI video startups—including HeyGen, Synthesia, Captions, and D-ID—each of which specializes in avatar-based video production.

Safety and Access Restrictions#

To prevent misuse, Google has implemented several protective measures. All avatars are bound to the account holder's identity, linked to their Google account, and embedded with invisible watermarking via SynthID. The company has also restricted access to adult users aged 18 and over in select geographies, a deliberate safeguard against potential deepfake or impersonation concerns. This contrasts with OpenAI's earlier approach to Sora, which allowed broader creative flexibility before the app was retired.