Alan Turing, the British mathematician and cryptogropher, is widely known as the “Father of Theoretical Computer Science and Artificial Intelligence”. Amongst his many accomplishments – such as breaking Germany’s Enigma Code – was the development of the Turing Test. The test was introduced by Turing’s 1950 paper “Computing Machinery and Intelligence,” in which he proposed a game wherein a computer and human players would play an imitation game.
In the game, which involves three players, involves Player C asking the other two a series of written questions and attempts to determine which of the other two players is a human and which one is a computer. If Player C cannot distinguish which one is which, then the computer can be said to fit the criteria of an “artificial intelligence”. And this past weekend, a computer program finally beat the test, in what experts are claiming to be the first time AI has legitimately fooled people into believing it’s human.
The event was known as the Turing Test 2014, and was held in partnership with RoboLaw, an organization that examines the regulation of robotic technologies. The machine that won the test is known as Eugene Goostman, a program that was developed in Russia in 2001 and goes under the character of a 13-year-old Ukrainian boy. In a series of chatroom-style conversations at the University of Reading’s School of Systems Engineering, the Goostman program managed to convince 33 percent of a team of judges that he was human.
This may sound modest, but that score placed his performance just over the 30 percent requirement that Alan Turing wrote he expected to see by the year 2000. Kevin Warwick, one of the organisers of the event at the Royal Society in London this weekend, was on hand for the test and monitored it rigorously. As Deputy chancellor for research at Coventry University, and considered by some to be the world’s first cyborg, Warwick knows a thing or two about human-computer relations
In a post-test interview, he explained how the test went down:
We stuck to the Turing test as designed by Alan Turing in his paper; we stuck as rigorously as possible to that… It’s quite a difficult task for the machine because it’s not just trying to show you that it’s human, but it’s trying to show you that it’s more human than the human it’s competing against.
For the sake of conducting the test, thirty judges had conversations with two different partners on a split screen—one human, one machine. After chatting for five minutes, they had to choose which one was the human. Five machines took part, but Eugene was the only one to pass, fooling one third of his interrogators. Warwick put Eugene’s success down to his ability to keep conversation flowing logically, but not with robotic perfection.
Eugene can initiate conversations, but won’t do so totally out of the blue, and answers factual questions more like a human. For example, some factual question elicited the all-too-human answer “I don’t know”, rather than an encyclopaedic-style answer where he simply stated cold, hard facts and descriptions. Eugene’s successful trickery is also likely helped by the fact he has a realistic persona. From the way he answered questions, it seemed apparent that he was in fact a teenager.
Some of the “hidden humans” competing against the bots were also teenagers as well, to provide a basis of comparison. As Warwick explained:
In the conversations it can be a bit ‘texty’ if you like, a bit short-form. There can be some colloquialisms, some modern-day nuances with references to pop music that you might not get so much of if you’re talking to a philosophy professor or something like that. It’s hip; it’s with-it.
Warwick conceded the teenage character could be easier for a computer to convincingly emulate, especially if you’re using adult interrogators who aren’t so familiar with youth culture. But this is consistent with what scientists and analysts predict about the development of AI, which is that as computers achieve greater and greater sophistication, they will be able to imitate human beings of greater intellectual and emotional development.
Naturally, there are plenty of people who criticize the Turing test for being an inaccurate way of testing machine intelligence, or of gauging this thing known as intelligence in general. The test is also controversial because of the tendency of interrogators to attribute human characteristics to what is often a very simple algorithm. This is unfortunate because chatbots are easy to trip up if the interrogator is even slightly suspicious.
For instance, chatbots have difficulty answering follow up questions and are easily thrown by non-sequiturs. In these cases, a human would either give a straight answer, or respond to by specifically asking what the heck the person posing the questions is talking about, then replying in context to the answer. There are also several versions of the test, each with its own rules and criteria of what constitutes success. And as Professor Warwick freely admitted:
Some will claim that the Test has already been passed. The words Turing Test have been applied to similar competitions around the world. However this event involved more simultaneous comparison tests than ever before, was independently verified and, crucially, the conversations were unrestricted. 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.
So what are the implications of this computing milestone? Is it a step in the direction of a massive explosion in learning and research, an age where computing intelligences vastly exceed human ones and are able to assist us in making countless ideas real? Or it is a step in the direction of a confused, sinister age, where the line between human beings and machines is non-existent, and no one can tell who or what the individual addressing them is anymore?
Difficult to say, but such is the nature of groundbreaking achievements. And as Warwick suggested, an AI like Eugene could be very helpful to human beings and address real social issues. For example, imagine an AI that is always hard at work on the other side of the cybercrime battle, locating “black-hat” hackers and cyber predators for law enforcement agencies. And what of assisting in research endeavors, helping human researchers to discover cures for disease, or design cheaper, cleaner, energy sources?
As always, what the future holds varies, depending on who you ask. But in the end, it really comes down to who is involved in making it a reality. So a little fear and optimism are perfectly understandable when something like this occurs, not to mention healthy.