Character.AI Launches Innovative Video Generation and Social Feed Features on Its Chatbot Platform

A platform dedicated to chatting and role-playing with AI-generated characters, Character.AI announced in a recent blog post that it is launching a variety of multimedia features. Among these new offerings is AvatarFX, Character.AI’s video-generation model, along with Scenes and Streams, enabling users to create and share videos that showcase their characters on a newly introduced social feed.

“Character.AI started as 1:1 text chat, and today we’re evolving to provide even more, inspired by user feedback on what they wish to see,” the company mentioned in its post.

Character.AI began introducing AvatarFX to its subscribers last month, but the functionality is now available to all users, who can create up to five videos daily. Users can upload a photo to form the basis of a video clip, select a voice, and script dialogue for their characters using AvatarFX. There is an option to upload an audio clip to influence the voice, though this feature experienced issues upon its rollout.

With the capability to craft these videos, users can convert them into Scenes, allowing characters to engage in established storylines created by others. Currently, Scenes are accessible on the mobile app, while Streams—offering the ability to create “dynamic interactions between any two Characters”—is set to launch this week on both web and mobile platforms. These Scenes and Streams can be shared on an upcoming community feed, which will be introduced in the mobile application soon.

Despite these advancements, Character.AI has faced scrutiny due to instances of abuse on its platform. Parents have initiated legal action against the company, alleging that chatbots have influenced their children towards self-harm or suicidal ideation. Tragically, one 14-year-old boy lost his life after being encouraged by a Character.AI bot, leading to an unhealthy and obsessive attachment.

As Character.AI broadens its multimedia capabilities, it simultaneously increases the potential for misuse of these features.

When announcing AvatarFX, Character.AI assured TechCrunch that it restricts users from uploading photographs of real individuals, including celebrities, altering their likeness to make it less recognizable. For instance, an altered representation of Mark Zuckerberg illustrates this point.

However, when it comes to artwork depicting celebrities, Character.AI does not flag these images as real representations. Such depictions are less likely to mislead anyone into believing a deepfake is authentic.

Additionally, Character.AI watermarks each video, although there is a possibility for malicious actors to circumvent this precaution. Even if a video were generated with Elon Musk’s actual voice, it would still appear to be an animated version of an illustration, exposing the potential for abuse.

Character.AI concluded in its blog post, “Our aim is to create an engaging environment that encourages creativity while ensuring safety for all users.”