Friday, October 17, 2014

Duplication of Human Intelligence with AIs

Daniel Bor presented an interesting comparison between computers and brains. He primarily differentiated between them by mentioning that computers rely on a binary code that only allows the computer to process serially and in a fashion that clearly determines the course of action based on previous decisions the computer makes. The brain, on the other hand, has thousands upon thousands of different routes (or neuron to neuron connections) that it can take that can make determining the outcome of a decision a very complex process. Daniel Bor used "parallelism" to describe the way the neurons in a brain process information because the human brain is able to apply vast conditional understanding of a question and enjoy the intricacies of why something isn't entirely a certain way or the exact opposite way. For instance, while a computer would answer questions on a yes or no basis such as whether or not it is raining outside, a brain would be able to determine the degree to which it is raining, the color of the sky, the scale at which the wind pulls on the body all while also processing the colors of the environment it is in. In short, just because you are trying to determine whether you should bring your umbrella with you, you can see the colors of your environment because your brain is processing information from your eyes at the same time. You may also look at how fast the rain is darkening a dry sidewalk to determine the rate of rainfall and in that way, you can merge your visual and memory knowledge of what the weather will be like in a way a computer would not be able to. This is just one of many ways that the brain is capable of running multiple operations at the same time that would contain too many variables for today's computers to process. A recent development occured among current computer programmers where they are trying to develop a Chatterbot that can pass the Turing test, or a test where the computer can fool a human being into believing that it is actually another human being interacting with him or her.

As a matter of fact, a Chatterbot recently passed the Turing test in the University of Reading in the UK. The Chatterbot, dubbed Eugene Goostman was able to convince a third of the humans that interacted with it that it is a 13 year old human boy. However, the age was used as a way to explain away that the computer kept saying that it knew everything that the human participants were asking it and kept giving erroneous answers. We can, however, accept Eugene as an accurate representation of a 13 year old boy because human children of that age present the same naive propensities. The clear next step is the development of more mature and refined machines capable of more complex parallel processing. This is but the beginning of the development of artificial intelligence in machines. The development of intelligent machines similar to Eugene will allow us to determine whether consciousness can be manufactured or whether it will be different even if the entire brain is duplicated, neuron by neuron, in a machine. 

"An AI milestone: Chatbot passes Turing Test by posing as 13-year-old boy." PCWorld. N.p., n.d. Web. 17 Oct. 2014. <http://www.pcworld.com/article/2361220/computer-said-to-pass-turing-test-by-posing-as-a-teenager.html>.
Bor, Daniel. The ravenous brain: how the new science of consciousness explains our insatiable search for meaning. New York: Basic Books, 2012. Print.​


No comments:

Post a Comment