How good are you at telling the difference between words written by a human and words written by a computer? Maybe after taking the Bot or Not test, you'll better understand how research publishers Springer and IEEE managed to miss gibberish papers.
Developed by two Ph.D. candidates at Australia's RMIT Melbourne -- Benjamin Laird and Oscar Schwartz (who's writing a thesis on whether computers can write poetry) -- the Web site is described as a "Turing test for poetry." Alan Turing developed the Turing test to determine whether intelligence is human or artificial.
The site has two modes -- a set test presented to attendees at the Digital Writers Festival in Melbourne from February 13-24; and Free Mode, which allows you to assess poems for as long as you like. The Web site will present a poem, and you have to guess whether it was written by a human poet or a computer program, such as jGnoetry and Ray Kurzweil's cybernetic poet.
The line between human and AI poets is becoming increasingly difficult to distinguish, as demonstrated by the site's leaderboards, which display the most human-like human poets and the most computer-like computer poets, as well as the most human-like computer poets and vice versa. For example, the computer-generated "A Wounded Deer Leaps Highest" is often mistaken by users as human poetry, while people seem to think Deanna Ferguson's "Cut Opinions" reads more like the work of a machine.
Computer poets are growing ever more sophisticated. Take Nathan Matias' Swift-Speare, an algorithm that can generate poetry in idiosyncratic styles, but only with the help of a human agent. According to Matias, though, a completely AI poet may not be far off. "I think I'll see a successful automated poet in my lifetime," he said. "It won't be easy: a poet is more than someone who makes poetry. Yet that doesn't rule out algorithms."