Can Machines Think?
“Can machines think? ”- A question heavily debated dating back to the days of Descartes. This idea has been explored and analyzed by scholars in many diverse fields of study for many, many years. Mathematician Alan Turing was a pioneer in unraveling this question. His philosophies found in his published paper “Computing Machinery and Intelligence” changed the ways in which we perceive and comprehend artificial intelligence. Turing’s most famous advance on the age-old question was the self-titled Turing test.
The test is quite simple; an isolated human investigator is given the task of distinguishing between a machine and a human based on the answers to the questions the interrogator presented. The test analyzes whether or not machines could be mistaken for humans by humans. If the machine is successful according to Turing it could be deduced that machines may be intelligent, display reason and bear the capacity to indeed think. In association with the Turing test, the CRUM hypothesis relates in many ways.
Computational-Representational Understanding of the Mind considers thinking to be made up of representational structures in the mind that are processed and represented by computational procedures. The main feature of CRUM is that the mind is composed of straightforward elements that work together to construct complex actions. In this way the Turing Test relates to CRUM as both programs or systems work towards exposing how the mind functions.
The Turing test attempts to prove that a machine is capable of interpreting and extracting complicated details via a structure of basic knowledge and algorithms, just like a human brain does. CRUM and Turing test both employ the theory that knowledge or action on the surface is released via a collection of information and refined implications In my personal view, I do believe that the Turing test is a sufficient control to determine if a computer is indeed thinking.
I believe that if a human is unable to differentiate from a machine and another human (as employed in the Turing test) that is a positive indication that the machine possesses the ability reason and communicate with another being that can indeed think. In September 2011 the program Cleverbot passed the Turing Test. 59. 3% of the simulations the investigator mistook the program for a human. In my personal experience I have had interesting conversations with Cleverbot, they are able to infer thoughts and respond with knowledge gathered from past conversations.
When deciphering if computers can think it is important to define what you believe thinking to be. Cleverbot can pass the Turin Test but there are certain types of questions that may stump the machine such as questions with the need for declarative memory instead of procedural memory. However in humans our knowledge is developed through experience, which is the same when Cleverbot is absorbing information from conversation to conversation. The Turin test is sufficient because it displays the computers ability to correspond with a being that has free thought. References
Embodied Mobile Robots Brian R. Duffy* and Gina Joue Media Lab Europe, Sugar House Lane, Bellevue, Dublin 8, Ireland Department of Computer Science The Trouble with the Turing Test Mark Halpern University College Dublin, Belfield The Best of Creative Computing Volume 1 Can Computers Thinkg? David H. Ahl ?An Invitation to Cognitive Science: Thinking? ? Daniel N. Osherson, Lila R. Gleitman http://blogs. discovermagazine. com/sciencenotfiction/2010/06/28/watson-fails-the-turing-test-but-just-might-pass-the-jeopardy-test/ http://stanford. edu/~rhorn/a/topic/phil/artclISSAFigure1. pdf
