Forget the Turing Test—give AI the F. Scott Fitzgerald Test instead

Written by Paul Sonderegger – Big data strategist, Oracle

Scott Fitzgerald pinned human intelligence on its tolerance of paradox. But what kind of artificial intelligence could pass his test?

In his 1936 essay “The Crack-Up,” Fitzgerald writes that “the test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function.” For example, he says you should “be able to see that things are hopeless and yet be determined to make them otherwise.”

He confesses he’s lost this ability—and as a result, himself.

Fitzgerald’s point is not that he needs a better model of the world, but that he needs many models and the freedom to switch among them. This is what allows us to forge ahead despite unexpected obstacles, conflicting priorities, or, in Fitzgerald’s case, hitting his forties and feeling like someone changed the rules of the game while he wasn’t looking.

Fitzgerald, having lost his ability to balance opposing ideas, falls into a drab, routinized existence. Every moment, from his morning routine to dinner with friends, becomes a forced act. He mimics the life of a successful literary man without actually living it.

Take a simple example. When a navigation app redirects stop-and-go traffic from the New Jersey Turnpike onto local roads in the town of Leonia, otherwise quiet neighborhoods become overrun with shortcut-seeking app-watchers. A compassionate human might weigh up the cost-benefit analysis of a shorter trip with the potential annoyance of hapless suburbanites. But a naïve AI, focused only on finding the fastest travel time, won’t. Local authorities are now plotting to fine non-residents caught driving through the area during rush-hour, even though they’re just following the directions on their smartphones.

Read More

Follow Us On Social Media:


  • No Related Posts

Leave a Reply