Deb Roy's home is full of ceiling-mounted video cameras that record nearly everything that happens within, including nearly every moment of his infant son's first years. Camera switches are as ubiquitous and unremarkable as light switches, only they're left on almost all the time.
Roy has dubbed his enterprise the Human Speechome Project. By the time it's done, he'll have 200,00 hours of video and audio that can be mined by software capable of recognizing patterns and themes -- of charting, in the greatest detail yet, the behavioral development of his child.
It's the prototype for a system where human analysts will look over the data, which in representational form look, at times, gentle ululations of color and motion, and at other times are the precise opposite: a vast, complex grid of checked and unchecked boxes. The analysts will be able to make and test hypotheses about behavior and meaning. Parents will know more about their children, at least in terms of sheer data, than ever before.
A few thoughts that come to mind: the tension between the visual representation and the underlying processes of learning and feeling and being a child. The gap between perception and reality, memory and media. Roy asks, "How does this transform us as people?"
He plays back with paternal pride a sonic collage of his son learning, over the course of months, to say "water." The boy's voice: "Gaga ...
googa ... gagoo ... wata ... googa ... gaga ... choo! ... goo ... wa
... choo ... wa ... chuh ... water." It is beautiful.
But I can't get the thought from my mind: what is the effect of permanent observation, of a domestic panopticon, of development tracked on the most complicated piece of graph paper ever? Will the early subjects establish a definition of "normal" that is, in fact, normal only to them? What happens when a child for the first time understands the nature of privacy?
