YouTube Launches AI Avatar Creation for Shorts, Featuring Voice Integration
YouTube has rolled out a new feature allowing users to create AI avatars that visually and audibly resemble them, specifically designed for use within YouTube Shorts. This functionality, which was previously previewed earlier this year, marks a significant step in personalizing short-form video content.
The new AI avatar creation tool enables users to generate photorealistic versions of themselves for YouTube Shorts, incorporating both visual likeness and voice.
Avatar Creation Process
The process for creating these personalized AI avatars is straightforward and accessible through both the main YouTube application and YouTube Create. Users are required to perform a "live selfie" recording, which involves capturing their face and voice by reading specific prompts. This recording session generates a photorealistic avatar, ready for integration into YouTube Shorts.
Each prompt-based generation sequence can last up to 8 seconds, with the flexibility to create multiple clips consecutively. The initial setup is a one-time requirement, though users have the option to retake the recording at any point to update their avatar's appearance.
This new capability expands Google's Veo models within YouTube Shorts. While a similar 'ingredients-to-video' feature, using uploaded pictures, has been available since last year, the voice integration aspect is a new and notable addition, enhancing the realism and personal touch of the avatars.
Safety and Usage Guidelines
YouTube has emphasized that this avatar feature is designed to offer users a secure method for including themselves in videos. The selfie video and voice data provided by users are exclusively utilized for avatar generation. Importantly, no other users can employ a created avatar to generate original Shorts.
Users maintain full control over their digital likeness, retaining the ability to delete their avatar at any time. For inactive avatars, YouTube will automatically initiate deletion after three years. However, existing videos that incorporate the avatar will persist until those specific clips are manually deleted by the user.
To ensure transparency and authenticity, all videos produced using an avatar will include watermarks and digital labels, such as SynthID and C2PA. These disclosures will clearly indicate that the content is AI-generated.
Availability and Access
The new AI avatar feature is rolling out globally, with the exception of Europe, beginning today. It is available to users aged 18 and older, with broader availability anticipated in the coming days. A prerequisite for creating an avatar is owning an existing YouTube channel.
To access the feature once it's available, users can follow these steps:
- Open the 'Create +' button in the YouTube app.
- Tap the Gemini spark icon.
- Select "Create video" in the top-left corner.
- Choose "Make a video with my avatar" to enter a prompt.
Alternatively, the feature can be accessed via the Remix menu by selecting Reimagine, then "Add me to this scene."