Featured Post

Hustler Fund Mobile app APK

Following the official launch of the Hustler’s Fund by President William Ruto on the 30th November 2022, Kenyans can now apply for the funds...

Tuesday, June 10, 2014

Eugene Goostman becomes the first AI to pass the Turing Test, convincing judges that he’s a 13-year-old boy

64 years after Alan Turing, the father of computer science, proposed a method of testing whether a machine has obtained human-level intelligence, a 13-year-old AI boy called Eugene Goostman has finally become the first artificial intelligence to pass the Turing Test. The test was carried out this weekend at the Royal Society in England, where Eugene managed to convince 33% of the judges that he was human (Turing set the threshold at 30% when he conceived the test in his seminal 1950 paper). Don’t worry, though, sentient computers aren’t about to take over the world: The Turing Test is actually rather flawed, and doesn’t really measure an AI’s capability for intelligent thought. Put it this way: There is a reason that very few top-end AI research groups are working on passing the Turing Test — because it would actually slow their development of a real human-level machine intelligence.
Back in 1950, Alan Turing wrote a paper called “Computing Machinery and Intelligence,” which attempted to tackle the tricky topic of whether machines are technically capable of showing intelligent behavior. In the paper, Turing proposed a test based on the Imitation Game — a party game where a man and a woman go into separate rooms, and then the other partygoers have to try and tell them apart by sending them questions and reading their typewritten responses. In the original Turing Test, one of the rooms contains a computer — and if it can fool the partygoers into thinking it’s a human, then it passes the test.
Diagram illustrating the Turing Test
Diagram illustrating the Turing Test
The modern-day Turing Test isn’t quite like that, though. In this case, Eugene Goostman — an AI developed by Vladimir Veselov, Eugene Demchenko, and Sergey Ulasen — is a chatterbot. Eugene is basically just a text box on a website: You type your message into the box, and then Eugene responds. At the event held at the Royal Society in London — organized by the University of Reading to celebrate the 60th anniversary of Turing’s death — a number of judges had a five-minute “conversation” with Eugene. 33% of the judges believed him to be human, passing the 30% threshold mandated by Turing, and thus becoming the first AI to pass the Turing Test.
By this point, you can probably tell that the Turing Test — especially the modern-day variation — is rather flawed. For a start, the judges already know ahead of time that computers are involved, and thus may be prejudiced to give a more optimistic (or cynical) response. We also don’t know what questions the judges asked (were they the right questions?) Presumably the test carried out at the Royal Society was double-blind (the judges didn’t know if they were talking to a human or a chatterbot), but the official press release doesn’t mention it, so maybe not (in which case, the results are worthless). But most of all, it’s important to note that the Turing Test doesn’t actually deal with actual intelligence — it’s only concerned with external behavior — how the machine acts — rather than what’s going on inside.
Alan Turing, with a Terminator eye that I photoshopped in. I'm not sure why.
Alan Turing, with a Terminator eye that I Photoshopped in. I’m not sure why.
In the case of Eugene Goostman, all he has to do is imitate a 13-year-old boy from Odessa, Ukraine. He doesn’t actually have to think about anything, or show any signs of sentience. In actuality, many of these chatterbots are just simple machines that follow a completely predetermined set of rules; they don’t learn, and they’re totally incapable of reacting to new situations that they haven’t been programmed for. They’re not really intelligent; they’re just mediocre conversationalists. [Read: How to create a mind, or die ]
So, while it’s nice to finally have the Turing Test sewn up after 64 years, you don’t have to worry about Eugene going all super Skynet sentient on us. There are plenty of R&D groups that are working on real human-level intelligence that could eventually lead to Judgment Day-like scenarios — but all Eugene does is provide a 13-year-old Ukranian boy for lonely people to talk to. (He’s online incidentally, if you want to see if he can fool you into thinking he’s actually a human boy, but the website is down right now, probably due to excessive load.)

No comments:

Post a Comment