When watching movies or reading books about the future a common, reoccurring element is artificial intelligence. Many futuristic stories portray a type of society in which humans either rely on, coexist with, or are threatened by ‘intelligent’ robots or computers. While stories like this used to be purely science fiction it seems that today artificial intelligence is becoming more of a reality. But can machines really think? How do we define thought? Will a machine ever be able to be considered truly intelligent?
A key signifier of artificial intelligence is the ability of a machine to pass a Turing test. To pass a Turing test basically means that a computer or program has the ability to fool a human into thinking it is not a computer or program. If a judge was asked to have a conversation with a person and a machine, both of which were placed in separate rooms where the judge couldn’t see them, and the judge was unable to distinguish between the machine and the human it could be said that, according to the Turing test, the machine displays intelligence.
Currently, there are many chatbots on the internet that try and pass this Turing test. I talked with three, and all of them failed(obviously, since I knew they were machines to begin with). While neither of the three were convincing enough for me to believe they were intelligent, they were all able to simulate conversation fairly effectively. ELIZA was probably the worst, it was very easy to confuse it with simple questions, and its speech wasn’t that convincing in general. Alice was more convincing in general but it still made some errors. At one point Alice made a weird totally unnatural statement using the word me instead of I and a strange sentence structure. I felt that JabberWacky was probably the best. I talked to the George bot, and even though it often made errors, its speech was the most natural sounding. I am not 100% sure if a bot will be made that can pass the Turing test in the next 10 years. I feel like it is definitely possible for it to happen if program developers work at it, I’m just not sure if it will be in 10 years.
When a program is finally developed that can pass the Turing test will this artificial intelligence be considered true intelligence? Many would say no. The argument being that there is a difference between the ability of a program to process information and output appropriate responses and actual human intelligence. While I do agree that there is a difference between imitating humans and real intelligence, I feel that it is possible for true artificial intelligence to exist. The human brain, as far as we know, basically just works from chemicals and electrical impulses. Therefore if a machine could be made that perfectly mimics a human brain, down to every last detail, then I feel that this machine would be able to learn, think, and understand things just as humans do, and would qualify as artificial intelligence.