Google Vids Now Lets You Star in Your Own AI Videos

Google has updated its video creation tool Vids to let users generate custom AI avatars from a selfie and a short voice recording, turning it into a more complete video production platform aimed squarely at business use cases like company updates and training material.

How the avatar feature works

Users can now create a personalized digital avatar that matches their likeness and voice, which can then be directed to appear in AI-generated videos built from text prompts and reference images through Google's Gemini Omni integration. The tool includes step-by-step editing capabilities along with the ability to swap backgrounds and adjust lighting after the fact, giving users a level of post-production control that previously would have required separate video editing software.

Built-in safeguards

Google has built in specific restrictions clearly aimed at limiting misuse: avatars can only be created using the account holder's own likeness, meaning users can't generate a video avatar of someone else without their direct involvement in the setup process. Access is also limited to users 18 and older in certain regions. Every AI-generated video carries an invisible SynthID watermark, Google's authentication technology designed to let the origin of AI-generated content be verified later, even after the video has been shared, downloaded or edited elsewhere.

Positioning against established competitors

The update puts Vids in more direct competition with companies that have built entire businesses around AI avatar video generation, including HeyGen, Synthesia, Captions and D-ID, several of which have raised significant venture funding specifically on the strength of this feature. Google's advantage is distribution: Vids is bundled into Google Workspace, meaning any organization already using Google's productivity suite gets access to AI avatar generation without adopting a new, separate tool or vendor relationship.

Why the timing matters

The update lands not long after OpenAI shut down its Sora video generation product in March, a move that left something of a vacuum in the consumer and business AI video space. Google stepping more aggressively into avatar-based video generation now, rather than earlier, suggests the company was watching how the market and the safety concerns around deepfake-style video tools played out before committing further engineering resources to its own version.

The business case Google is making

By framing the feature around company updates and training videos rather than general entertainment content, Google is deliberately steering Vids toward a use case where the safety risks are lower and the commercial value is clearer: a company making an internal training video with an employee's own avatar carries far less risk of the kind of consent and identity-fraud problems that have dogged more open-ended AI avatar tools.

Whether the safeguards Google has built in prove durable against determined misuse will be the real test, but for now the company is betting that tying avatar creation strictly to verified self-likeness, combined with SynthID watermarking, gives it a defensible position in a feature category that has otherwise drawn persistent scrutiny over consent and authenticity.

For businesses already paying for Google Workspace, the calculus is straightforward: rather than evaluating and onboarding a separate avatar-video vendor, teams can experiment with the feature inside a tool they're already using for documents and slides, lowering the barrier to trying AI-generated video for internal communications long before any organization-wide commitment to the technology.