A recent article in the NY Times (I’m afraid a subscription is required to read it) discusses a new I.B.M. supercomputer called Watson. Watson is designed to answer questions, and not just any questions, but the sort of tricky questions used in the popular game show Jeopardy.
Apparently Watson is quite good at it. He won several rounds of Jeopardy in private test runs. Furthermore, the folks at Jeopardy have agreed to have Watson on as a contestant for an episode of the show later in the year. That should be pretty freakin’ amazing…
This story comes on the heels of an essay I have just finished writing on free will (soon to be posted here). In this essay, I discuss how the operation of the mind, like the computer, is based on relatively simple physical components which, when combined in clever ways, give rise to the full complexities of thought and personality.
The Watson story, I believe, is another step toward convincing religious dualists that complex thought is, indeed, possible using wholly natural processes, and that no “soul” or “spirit” is needed.
Human contestants who competed against Watson were obviously impressed by his performance, and fell quickly into treating Watson as a real personality, referring to the machine as “him”, and making remarks about its game play (e.g., “He plays to win”, “He’s really not messing around”). This reminds me of the Turing test, which suggests that a machine can truly be regarded as intelligent if humans cannot distinguish its behavior from that of their fellow humans. There have been a lot of objections to this idea over the years, but I think it serves as a decent enough rough test of intelligence. Watson demonstrates that this test is getting closer to being passed.
Of course, Watson is not a polymath: “his” abilities, while remarkably deep, are fairly narrow, and even then he requires a “brain” that fills a room. But it’s a sign of progress nonetheless.