Ostensibly a 13-year-old boy from the Ukraine, the program was able to pass the Turing Test, which states that a computer could be considered to be "thinking" if it could fool 30 percent of researchers during a five-minute text conversation. First dictated in 1950 by computer pioneer Alan Turing, the test is considered the preeminent benchmark for computational philosophy and artificial intelligence. The Russian-made program, tested by the Royal Society in England, fooled 33 percent of its interrogators.
As reported by the Independent UK, in regards to his success Mr. Goostman stated, “I feel about beating the turing test in quite convenient way. Nothing original."
The success of the test brings about many questions, including many regarding the safety of computer users when dealing with possible cybercriminals. Fortunately the Goostman program has not evolved to a stage of teenage mischief...yet.
UPDATE: The validity of this article has been proven wrong. My apologies, that's what I get for trusting the corporate media and their wannabe-robot minds.
33% !? There are two outcomes; the enity taking the test is a human or it is an AI. Given that we would expect that flipping a coin would on average give the right answer 50% of the time! The only place I can find references to turing saying 30% is in articles about this same chatbot. Some of which date back to 2012: http://www.newscientist.com/blogs/onepercent/2012/06/bot-with-boyish-personality-wi.html This is a non-story being trotted out for at least the second time.
ReplyDeleteThis study may say more about random people's estimation of 13 year old boys than anything else. TechDirt also provides a critical perspective on the report:
Deletehttps://www.techdirt.com/articles/20140609/07284327524/no-supercomputer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml
My apologies for the inaccurate post. AdminX's attached article is wholly correct.
ReplyDeletenice post
ReplyDelete