Google Updates Flow AI Tool with Deepfake Avatar Feature
Original: Google Makes It Easy to Deepfake Yourself
Why This Matters
Marks Google's entry into creative AI tools, competing with OpenAI's video generation capabilities
Google enhanced its Flow AI video creation tool with avatar feature that lets users create deepfake versions of themselves. The update includes Omni Flash model for improved video generation and automated workflows for creators.
Google announced updates to Flow, its experimental AI video and image generation tool, at its I/O developer conference. The key addition is an avatar feature allowing users to create digital clones of themselves by scanning a QR code and recording their face and voice. Users can then insert themselves into AI-generated videos using the new Omni Flash model, which succeeds the Veo model. Elias Roman, VP of product management at Google Labs, says the feature targets creators who want to appear in content without filming themselves. The avatars are also available through Gemini app and YouTube. Flow includes automated workflows and custom instructions for video generation. Unlike OpenAI's defunct Sora app, Google's avatars only allow users to create versions of themselves, not others.