As our world becomes integrated by smarter and smarter machines, interacting with them is going to become more and more important. Finding common perceptual grounds — visual, audible or haptic grammars that humans find compelling and computers find richly accessible to processing and manipulating — is one of the hard problems of robotics and artificial intelligence.
Suwappu, a prototype augmented-reality toy system consisting of blocky little figurines who live in their own virtual environments, is one attempt at solving this problem. A project of design shop Dentsu London, Suwappu explore a new approach to augmented reality play, in which an app "reads" the figurines' simple faces and color patterns much the way we do.
Whether they ultimately reach the market or not (as the video shows, the project remains design fiction at this point), Dentsu's Suwappu offer a gesture in the direction of human-robot interoperability. Endearing and compelling, the figures' faces and color schemes signal character to us; at the same time, the computers find them susceptible to processing. Suwappu suggest that humans and machines already share a visual vocabulary, a kind of machine-human graphical creole.
Roboticists and animators already have the "uncanny valley," that threshold of verisimilitude where nearly real androids and animated characters unsettle. Perhaps there’s an analogous boundary zone — call it an "isthmus of intelligibility" — where machines edge ever closer to seeing things as we see them. With the common perceptual grammar of humans and machines — the lingua franca of the uncanny valley — growing richer, we’re inching out across that isthmus already.
[Hat tip: Tim Maly at Quiet Babylon]