aNewDomain — By now you’ve heard that a chatbot, posing as a somewhat-snotty 13-year-old Ukrainian boy named Eugene Goostman, passed the 65-year-old Turing Test.
That means, by definition, that it managed to fool more than 30 percent of humans into thinking it was a human and not a machine (by using an attitude and an accent). This was a first, researchers said, and the media by and large jumped excitedly on the hype-ish news bandwagon. Most reporters fully ignored other reported Turing Test “wins” since 1990. Few questioned the wisdom of using a trick to outwit the test.
Using the ruse of a Ukranian teen with an attitude is clever. The bot can pretend to not understand the questions, and it can respond snottily or with bad English grammar. But think. Couldn’t anyone create a bot that was, say, an angry severely schizophrenic person — and similarly pass the test? Ask it a question and it screams nonsense at you. Cake.
Could that bot pass as human to 30 percent of people in chat-based questions? Why, sure!
It really is that subjective.
On news that the bot passed the test — successfully fooling 30 percent of human interrogators at the University of Reading into thinking it was a human, via a series of five-minute keyboard “conversations” — the BBC and other media outlets crowed that a major milestone had been reached.
Check out the video of the chat-based test, below.
Video: Beyond Reason Designz
Researchers celebrated at the June 7 “victory” over Turing’s 65-year-old AI challenge. Said the University of Reading’s Kevin Warwick:
The words … Turing test … have been applied to similar competitions around the world. However, this event involved the most simultaneous comparison tests than ever before, was independently verified and, crucially, the conversations were unrestricted … (but) a true Turing test does not set the questions or topics prior to the conversations. We are therefore proud to declare that Alan Turing’s test was passed for the first time on Saturday (June 7, 2014).”
In the BBC, AI expert Lord Sharkey called the June 7 “Eugene bot feat ” a great achievement for Eugene.
It was a very clever ruse to pretend to be a 13-year-old Ukranian boy, which would constrain the conversation. But these competitions are really great to push developments.”
Well, exactly. And it was impressive, certainly, for the fake Eugene Goostman.
Amidst the misplaced hype, New Scientist‘s Celeste Biever was a rare voice of reason. In her piece, she said:
Would you consider a teenager the pinnacle of human intelligence? Probably not. This gets at something that is both clever – but also limiting – about Eugene. Passing the Turing test is touted as being the moment that machines have got us, since they can perfectly imitate people. But it doesn’t address which people the machine must imitate in order to pass, raising the question of whether a 13-year-old boy really counts. When I met (Eugene chat bot creator) Veselov in 2012, he said … ‘thirteen years old is not too old to know everything and not too young to know nothing.’ In other words, it’s old enough to hold a conversation, but much easier to mimic, than, say, Richard Dawkins, or Stephen Hawking. By aiming low, Eugene succeeds at being a realistic human character – but is he really all that smart?”
For more, see the full article here.
Sounds like a gimmick — because it is.
For aNewDomain, I’m David Michaelis.
Based in Australia, David Michaelis is a world-renowned international journalist and founder of Link Tv. At aNewDomain.net, he covers the global beat, focusing on politics and other international topics of note for our readers in a variety of forums. Email him at DavidMc@aNewDomain.net.
