Developer Pranav Mistry, Lead Project Neon at Samsung, posted a jubilant tweet about his willingness to introduce the new technology to the world. It received the working title "artificial people" and, apparently, is the most advanced deepfake in the world today. At least, it is in this vein that Mistry describes it, but with a caveat - the project is intended for good purposes.
According to Mistry, “artificial people” are now full-fledged avatars of real people in the digital space, but are no longer attached to them. It is enough to conduct an initial scan, build a personal model, after which it will be able to act independently. For example, move in the frame, dance, sing, including in exotic languages unknown to the "owner" of the avatar. In short, live in a digital space.
It is not yet entirely clear whether the image is generated by a computer, or if it is a variation of video motion capture. In the example laid out, "digital people" look convincing, but what about their ability to act in a real environment? Will the AI, for example, be able to control such an avatar to tell a joke in a way that really makes the audience laugh, using realistic facial expressions and gestures?
When asked why these "digital people" are needed, Samsung has long prepared the answer - this is the future. And the nearest, since in the next decade all kinds of services will acquire not only a human "face", but also a full-fledged "body", with a voice and personal settings. They will talk to us from the screens of gadgets, provide assistance in the form of holograms on the ground, at important objects, and move into virtual reality. This is not a deepfake, not content fake, but a much deeper technology.