October 16, 2014
As computing power increases and man-made computers continue to knock down humans in competitions of the intellect, the world of science fiction continues to blend with present-day reality. Many futurists predict that we will be able to understand and duplicate human brain function within the next few decades, an important step that may lead to the creation of a computer that is truly smarter than a human, with full creative capacities, emotions, and more.
One important marker of machine intelligence is the Turing Test. The test was first conceived by Alan Turing in 1950 as a way to evaluate computer intelligence, and possibly as a sign of real thinking, a function that many hold to be uniquely human.
The test is simple in design — one may even call it elegant. A human judge sits in one room and engages in two conversations, one with a human, the other, a computer. The conversation carries on solely through text. The judge’s goal is to distinguish the human from the machine.
As computing power progresses, the ability of machines to successfully mimic human speech patterns does as well. Many believe that the day a machine successfully passes the Turing Test will be a milestone, perhaps a sign of the last days of human supremacy in the realm of thought.
There are arguments, however, that predict otherwise. John Searle, in a landmark argument in the philosophy of mind, claimed that mere symbol manipulation does not imply thought. That is, that responding to stimuli along a pre-determined route does not require understanding.
However, consciousness does, and so, when a person speaks, they not only command symbols, they understand meanings. In the case of a machine, which only exchanges one symbol for another, even highly complex interactions are not enough to qualify as thought. So the argument goes.
Searle’s objection aside, passing the Turing Test is widely regarded as a sign of consciousness. At the very least, it’s a milestone in computing, as the number of connections required to simulate human conversations is staggering. And, until recently, no machine had successfully passed the test.
That changed on June 7, when a Russian chatter bot named Eugene Goostman was declared a human by 33 percent of judges. However, the announcement brought instant international outcry. As it turns out, the bot made the very clever (humanly clever, can we say?) move of telling judges it was a Ukrainian national speaking English as a second language. Many critics claim that this move should not be allowed, as it changes the parameters by which the conversation is judged and throws too much uncertainty on the outcome. Furthermore, the bot said that it was only 13 years old. This move was also challenged, as traditional Turing subjects are assumed to be of adult age. Then, there’s the low percentage. While 33 percent isn’t insignificant, it’s certainly not overwhelming. Perhaps, for a milestone as significant as the Turing Test, our standards should be higher.
Written by Ohad Mark Stoller of Fueled. Fueled.com is an award-winning mobile app design and development house based in New York, Chicago and London. At Fueled, we don't just build apps; with teams of designers, developers and strategists, we create visually stunning products that redefine the technical boundaries of today's mobile development standards. We've built award-winning iPhone, iPad and Android apps used by millions of people for clients ranging from Fortune 100 companies to up and coming startups including Barney's, Coca Cola, UrbanDaddy, JackThreads and MTV. We hold ourselves to the highest standard of usability, stability and design in every project that we touch.