YouTube has launched new AI-driven features within YouTube Shorts that allow creators to create clips using digital avatars that look and sound like themselves. The company had previously announced this feature, and is now officially launching it on the platform.
The setup process requires a “live selfie” recording, where the user records their face and voice while reading a set of phrases.
The system then creates a realistic avatar that can be used within Shorts clips. A single clip can be up to eight seconds long, with the ability to merge multiple clips together. It is also sufficient to create an avatar only once, with the ability to update it later.
YouTube confirms that the user will only be able to use his own avatar, and the use of other people’s avatars will not be allowed.
Avatars are also automatically deleted from the company’s servers after three years of non-use. After setting up your avatar, clips can be created via text commands using Google’s Veo text-to-video technology.
The feature began rolling out globally starting April 9, excluding Europe, via the main YouTube app and the YouTube Create app. The feature is available to users aged 18 years and over who have channels on the platform.
On the other hand, this move raises concerns regarding deepfakes, as these tools may make it difficult to distinguish between real and artificial intelligence-generated content. This is why YouTube confirms that it will clearly label content created using artificial intelligence, and also requires content creators to disclose any modified or artificial materials.
In addition, the company has introduced a deepfake detection tool aimed at politicians and journalists, in an attempt to boost trust on the platform, although the biggest challenge of distinguishing between the real and the fake will remain as these technologies develop.
