You know about artificial intelligence, powerful computer software that can duplicate certain feats of human intelligence.

We've been told repeatedly for the last decade or so to ratchet back our expectations about the potential for artificial intelligence. Yes, it's AI when the creatures you're shooting at in a video game learn how to escape your blasts, and when Amazon.com tells you that if you like The Great Gatsby you're the sort of person who might want to buy a Weber barbecue grill.

Big deal.

But I've just heard about a new instance of artificial intelligence in the cultural realm that seems like something out of The Jetsons - or, more accurately, Sleeper. A company called Polyphonic HMI has engineered a piece of software called Hit Song Science.

And when Polyphonic HMI plays a piece of music on a computer loaded with the Hit Song Science software - any piece of music, from any genre - they say they can tell pretty accurately if that song has the potential to become a pop hit.

Programmed into the software are the salient acoustical data - the speeds, rhythms, harmonies, and keys -- of every successful pop song from the last 30 years.

The company's engineers apparently first realized that their software really, really worked two years ago, when they analyzed Norah Jones' debut album ''Come Away With Me''

The Hit Song Science software predicted that eight of the 14 songs on the album could become hits.

Which was ridiculous. The smart money in the music business was most definitely not on Norah Jones becoming a big pop star. In addition to her own complicated, low-key tunes, she sang covers of Hank Williams and Hoagey Carmichael.

The music was too low-key to be really commercial, too interesting and sui generis.

But over the next year Norah Jones' debut album sold eight million copies. She also won five Grammy Awards.

So now, of course, the record companies are getting Polyphonic HMI religion. Their marketers are using Hit Song Science to help decide which songs on a particular album to market most energetically, and producers are using it to help shape works in progress while they're making the albums.

On the one hand, given that newfangled digital technology, in the form of free song-downloading software, is a significant source of the music industry's troubles, it seems poetically just that they're using an even newer-fangled digital technology as they struggle to survive. On the other hand, does any sector of the entertainment industry really need encouragement to produce new product that resembles successful old product?

We thought we'd see what it was like to take a ride on this high-tech bandwagon. So we asked some musicians we know if we could submit some of their songs to Polyphonic HMI for analysis. It cost us $44.99 apiece.

And according to Hit Song Science the song, One Tomorrow, by our friends The Grocery Concern is exactly average in its hit potential: on a 1 to 10 scale, it scored just over 5. Alas, 7 and up is the surefire hit zone.

The actual hits in the Hit Song Science database that most closely resemble One Tomorrow are Angel by Sarah McLachlan and Only Time by Enya.

This is the most interesting part of Hit Song Science to me - finding those underlying similarities between apparently rather dissimilar music. According to the New York Times magazine, the HSS software says that the music of the band U-2 is in fundamental, scientifically measurable ways similar to the music of Beethoven.

Which, now that I know it, seems weirdly, intuitively right. Or at least explains why I happen to be a big fan of both Achtung Baby and Ode to Joy.