Cracking the Turing Test

In 1950, the mathematician Alan Turing wrote a paper entitled "Computing Machinery and Intelligence." The paper began with the simple and now legendary phrase, "I propose to consider the question, ‘Can machines think?'" From this paper came the Turing test — in many ways the definitive way to answer this rather vague question. The basic idea is simple, although (predictably) academics disagree on the details. You get a volunteer to converse with a number of other volunteers and one computer. If they can't tell the difference between the computer and the humans, then the computer is said to have passed the Turing test. This test is a major benchmark in artificial intelligence and no machine has yet convincingly passed the test.

So imagine the excitement when this video was released about a year ago. At a Microsoft conference, Peter Molyneux, head of Lionhead Studios, unveiled a new project. A volunteer named Claire interacts with a virtual ten-year-old boy called Milo in a rather impressively realistic way. Milo can apparently recognize faces, react to speech, and even get evasive when asked if he's done his homework. Cue a deluge of responses from the blogosphere. Surely it's not real, its trick, its just hype, etc. . . .

So is it real? Well, as usual that depends on what your definition of real is. A more recent presentation at TED seems to imply that the technology works by recognizing subtle changes and inflexions in the voice and then reacts to those changes in a way that makes sense. For example, a calming voice is useful in the game when Milo is feeling down. The program also recognizes certain words, making for a combined effect which overall I must admit is pretty realistic.

What the technology does is allow a new dimension for computer games and for enthusiasts that's bound to be a good thing. Unfortunately, it doesn't come any closer to cracking the Turing test. Indeed, Molyneux admitted in a recent interview that if they had cracked it, applying it to a computer game would be the last thing he'd use it for. Personally, I won't be buying a copy of Milo and Kate when it's released (I waste enough of my time on computer games as it is!), but I think it is safe to say that we can expect some new and exciting things from this corner of the gaming industry in years to come.

Further Reading:

Turing, A. "Computing Machinery and Intelligence." Journal of the Mind Association 59, 433–460 (1950).

NB. Kotaku and other sources have reported that the Project Milo has been canceled.