“I propose to consider the question, ‘Can machines think?’” With this rousing opening line, Alan Turing began his seminal 1950 paper on “Computing Machinery and Intelligence” and ushered in a futurist paradise (or nightmare, depending on your outlook) where the line between humans and their mechanical creations is rendered ever more blurred by the bounding advance of artificial intelligence.

 

By re-defining the esoteric boundaries of his original question and instead focusing on whether or not machines could “do as thinking entities could do,” Turing transported the debate on artificial intelligence from the heady realms of philosophy to the spheres of practicality, generating an elegant and relatively simple test to illuminate the “humanness” of AI responses.

 

In his paper, Turing proposed an “imitation game” of sorts, one where “a computer convinces a sufficient number of interrogators into believing that it is not a machine but rather is a human.” Human judges engage in text-based conversations with test participants, endeavoring to ascertain whether or not the subject is human. This thought experiment quickly became known as the Turing Test.

 

On Saturday, June 7, the University of Reading announced that a software program known as Eugene Goostman had successfully passed a Turing Test competition, the first AI program in history to do so. By convincing 33 percent of the judges that it was a 13-year-old boy from Ukraine (traditional interpretations of the Turing Test demand a hoodwink rate of one-third of the judges for successful completion), the clever little chatbot won the competition and immediately generated a flurry of controversy in the AI community.

 

Objections to Eugene’s victory are flying hot and fast, with AI luminaries such as Ray Kurzweil (the author of “The Singularity Is Near,” which your correspondent opines is one trippy book indeed) and Netscape co-founder Marc Andreessen happily eviscerating the cheeky bot’s performance. Points of issue include the make-up of the test (text conversations were limited to five minutes) and the relatively low pass rate (some argue that the one-third requirement confuses a futurist prediction by Turing with an attempt to set the spectrum of the test). Some objectors have accused Eugene’s developers, a Russian and a Ukrainian, of gaming the test by masquerading the chatbot as an adolescent who spoke English as a second language, which gave judges the liberty of discounting grammatical errors and confusing non-sequiturs as products of youth and a language barrier.

 

Much of the uproar has deeper roots in whether or not the Turing Test is even a relevant measurement of AI’s advance at all. Computer programs have allegedly passed the test before, though Reading’s organizers argue their version of the test is a better representation. Humans are easily fooled, which is why programs like Eugene can use offbeat humor to disguise an actual inability to understand certain questions. In a post for the New Yorker, cognitive scientist Gary Marcus pointed out that Turing’s true vision of “a flexible, general-purpose intelligence of the sort that human beings have” has essentially gone unrealized in an era of highly specific AI programs devoted to highly specific tasks. The test, then, is a quaint relic rather than a viable measurement of progress.

 

Two things seem clear in all of this. First, Eugene is no Skynet. Rather, it is a cleverly designed pattern-seeking bot capable of bluffing its way towards an illusion of humanness but unable to sustain the farce for very long (for an excellent illustration, check out this Mashable article or chat with Eugene yourself!) Second, the debate on artificial intelligence remains as confusingly entrancing as ever. The terminology used in dissecting Eugene’s so-called achievement is rife with words like “deceitful”, “deflection,” and “misdirection,” imbuing the very bot writers seek to eviscerate with an almost human aura of mischief and sleight-of-hand. Perhaps this is the real genius of programs such as Eugene – in our efforts to ridicule their achievements, we allow them space on our philosophical playing field.

 

What do you think of Eugene Goostman? Are you impressed, or is this all smoke and mirrors? Let’s talk here, or find me on Twitter @aa_murph