What to know

  • Character.AI has launched AvatarFX, an AI video model in closed beta, to create lifelike animated chatbots from images and text.
  • AvatarFX can animate existing images, generating photorealistic videos with synchronized facial expressions, gestures, and speech.
  • The tool is currently available to Character.AI+ subscribers, with broader access planned soon and a public waitlist open.
  • There are ongoing concerns about misuse, including deepfakes and emotional manipulation, prompting Character.AI to enhance safety features.

Character.AI has unveiled AvatarFX, a new AI video model designed to create highly realistic animated chatbots. This tool is currently in closed beta and allows you to generate animated videos of platform characters using only text and images.

Unlike many other AI video tools, AvatarFX can animate existing images, which means you can upload a single photo and turn it into a video where the character displays lifelike facial expressions, hand gestures, body movements, and even synchronized speech or singing.

AvatarFX is powered by a flow-based diffusion transformer pipeline and a custom text-to-speech engine. This combination enables the model to handle long video clips, multi-speaker dialogues, and user-defined keyframes, giving you precise control over how your animated scenes unfold.

The tool supports a wide range of animation styles, from photorealistic humans to 2D cartoon animals, making it versatile for various creative needs.

Currently, AvatarFX is accessible to Character.AI+ subscribers through the platform’s experimental Lab. The company plans to integrate it into the main app in the near future, and a public waitlist is available for those interested in early access on both web and mobile platforms.

However, the launch of AvatarFX has raised concerns about potential misuse, particularly the risk of creating deepfakes and the possibility of emotional manipulation by AI-generated avatars.

Character.AI is already facing lawsuits related to harmful content produced by its chatbots and has responded by introducing new safety features, including parental controls. The effectiveness of these measures depends on how users implement them, highlighting the need for ongoing vigilance when interacting with advanced AI technologies.

Via: TechCrunch