Copyright Talk

Raptors - those magnificent birds of prey that include hawks, owls, and eagles - have taught aeronautic engineers a lot about flight maneuverability at high speeds. Our modern fighter jets owe a great deal to the elegant design of the raptors. The jets are not so elegant as the birds, but they are a ton more maneuverable thanks to the fact that hawks do not have a copyright or patent on their bodies.

I watched a show last night on PBS that discussed how raptors are designed to do the things that they do, and showing the adaptations that have been made to fighter planes based on observations of these birds. Fascinating. Anyway, the show also mentioned something which owls and eagles are particularly known for - their vision acuity. Owls and eagles are able to spot and follow prey from great distances, in part because their eyesight is much sharper than that of the human eye. Apparently, this is because the birds have a “dimple” on their retina, effectively giving them two fields of vision at a time, one of which acts basically as a telephoto lens. But they have other adaptations as well, including the ability to see a broader spectrum of light, and the ability to “stabilize” the object they are stalking while moving in on it at a high rate of speed, so as to prevent losing sight of it altogether.

In other words, the raptors have sight that allows them to do multiple things at once, all of which are necessary to find and catch small prey.

So it occurred to me that perhaps raptors can teach us not only how to fly better, but perhaps how to do other things better as well, such as monitor computer data for copyright violations. The greatest concern raised about currently developing technology used by youtube and others, is that it has only the ability to spot similarities in data, but not to spot actual copyright violations versus legitimate uses of material, such as those which fall under fair use. Like the vision of raptors, any technology designed to automatically detect copyright violations will have to be able to do several different things at once - determine similarity, determine context, and determine amount. Anything that fills less than all three needs, simply falls short, and will inevitably lead to errors.