Charature.AI, a platform for AI-generated characters and chat and role-playing, announced in a blog post Monday that it is rolling out a multimedia feature slate. These features include video generation models from Avatarfx, Character.ai, as well as scenes and streams. This allows users to create videos featuring characters and share them in their new social feeds.
“character.ai started out as a 1:1 text chat, but today it has evolved to do more, inspired by what users said. [us] They want to see it on the platform,” the company wrote in a blog post.
Character.ai began rolling out Avatarfx to subscribers last month, and now everyone can create up to five videos each day. When creating videos using AvatarFX, users upload photos to act as the basis for video clips, select audio and write character dialogue.
There is an option to upload audio clips to notify you of the sound of the audio, but this feature wasn’t working well enough to test on rollout.
Users can turn these videos into scenes. In the scene, characters can slide into a pre-populated storyline created by others. While scenes are currently available in mobile apps, Streams, which allows users to create “dynamic moments between two characters,” will be appearing on both the web and Mobile this week. These scenes and streams can be shared with the new community feed that will soon be coming on the mobile app.
Character.ai has a track record of abuse on the platform. The parents filed a lawsuit against the company, claiming that the chatbot tried to persuade their children to self-harm, commit suicide, or kill their parents. One 14-year-old boy died of suicide after being encouraged by a character.
As Character.ai expands its multimedia products, the possibility of these products being abused will also increase.
As Character.ai told TechCrunch when it announced Avatarfx, the platform blocks users from uploading photos of real people, whether celebrities or not, and obscures their portraits with lesser recognition.
For example, the Uncanny Valley version of Mark Zuckerberg’s Charature.ai.
However, when it comes to artwork depicting celebrities, characters don’t flag the image as representing real people, but this kind of portrayal makes it less likely that deepfakes will deceive someone to believe they are real.
Furthermore, the character’s watermark is on each video, but it is possible for bad actors to navigate their safety measures.
Here is an example of Deepfark’s attempt based on Elon Musk’s illustrations:
Even if this video was generated with Elon Musk’s actual voice, it is still relatively clear that this is an animated version of the illustration, but the possibility of abuse is clear.
“Our goal is to provide everyone with an attractive space that promotes creativity while maintaining a safe environment,” says Charition.ai in a blog post.
Source link
