danbroussard

Artificial vision gets sharper all the time. Artificial walking has made great robot strides. But artificial intelligence is brain-dead. Why? Because while researchers have built awesome technology, they've failed to grapple with philosophy.

Early computer scientists like Alan Turing and John von Neumann attempted to spin logic circuitry into thinking machines. So they designed their creations to excel at tasks they thought embodied the heights of human intellect: calculating sums, analyzing geometry, and playing chess. In 1958, a computer beat a human chess player — albeit an inexperienced one — for the first time. Buoyed by this success, researchers embarked on a long attempt to invent a machine that could (a) talk just like Turing, (b) walk around in a robot body looking at stuff, picking it up, and using it neatly, (c) read newspapers and maybe correct the editors, and (d) program its own successors.

The notion of intelligent machines inspired a blizzard of books and movies, but practical returns were meager. Over the years, computers failed one commonsense task after another: manipulating unfamiliar objects, understanding natural language, distinguishing a dog from a cat. Despite steady advances in hardware, no machine could think as far as to laugh at a pratfall or make a bad pun.

In pursuing human-style intelligence, the geeks blundered into the deepest, densest, darkest thickets of metaphysics: consciousness, cognition, perception, self-awareness, and how "we" manage to "know" what we know. It turns out that activities like playing chess — things that require sorting and searching — are relatively easy to program, whereas tasks that require some understanding of the world at large, like doing the laundry, are unbearably complex. The metaphysical issues around AI are at a standstill, mainly because metaphysics is old and canny and doesn't move forward in the linear manner of technology. Researchers, their grand illusions and ambitions dashed, fell into a long "AI winter" of shrunken budgets and general indifference.

Nowadays, Google "knows" pretty much anything you ask it. But its insanely fast and powerful work is modestly described as data-mining, not thinking. That vast, globe-spanning, superpowerful, ultrawealthy Web spider has yet to awaken and declare, "I am Google."

But if it starts writing philosophy, all bets are off.

posted by Daniel Webster Broussard  # 9:42 PM