"Chess Is Too Easy" by Selmer Bringsjord, in Technology Review (Mar.–Apr., 1998), Massachusetts Institute of Technology, Bldg. W59, Cambridge, Mass. 02139.

When IBM’s Deep Blue bested world chess champion Gary Kasparov last year, some scientists hailed the victory as a landmark on the way to creation of a machine with intelligence equal to the human sort. Bringsjord, who teaches logic and artificial intelligence (AI) at Rensselaer Polytechnic Institute, argues that while computers may regularly checkmate human grand masters one day, they will never achieve intellectual parity with their creators.

Deep Blue’s triumph was a victory for proponents of so-called strong AI, who believe that all human thought can be broken down into a series of mathematical operations. If that sounds impossible, so, until recently, did formidable chess-playing computers—at least to some experts. In his 1992 book What Computers Still Can’t Do, Hubert Dreyfus, a philosophy professor at the University of California, Berkeley, said that such machines would forever remain science fiction. Yet chess, Bringsjord points out, theoretically can be reduced to a series of mathematical operations. The true test of computer intelligence, he argues, lies in something far more elusive: the ability to create.

A genuinely intelligent computer, for example, would be able to write fiction that is rich in language, plot, and characterization. For the last seven years, Bringsjord has been working to build "a formidable artificial author of short short stories." The latest result, he says, is a machine named Brutus.I, which can compose very short stories, provided they "are based on the notion of betrayal (as well as self-deception, evil, and to some extent voyeurism)." This feat was made possible because Bringsjord and a colleague were able to devise a formal mathematical definition of betrayal and implant it in the machine. But Brutus.I gets writer’s block when it comes to other great literary themes, such as revenge and unrequited love.

Bringsjord’s 10-year quest to construct a "silicon Hemingway" has three years left, he notes, but it already "seems pretty clear that computers will never best human storytellers in even a short short story competition." For a machine to tell a "truly compelling story," he points out, it would have to understand the characters’ "inner lives"—and that would require not just swift calculation à la Deep Blue but the ability "to think experientially," mixing memory and perception as an artist does. The chess champs of the future may have reason to worry, but John Updike and his successors do not.