Discuss:Mathematics, Logic, Philosophy:AI
From da Vinci Concept
Contents |
AI
henrik - Tue Dec 21 8:00:01 2004
Is it possible to faithfully emulate a human mind on a Turing machine, or some approximate physical implementation of a Turing machine such as a computer?
--henrik Tue Dec 21 8:00:01 2004
macrocosm - Tue Dec 21 8:38:47 2004
What is meant by "faithfully"? Recent breakthroughs in physics indicate that there is more to us as human beings than our organic components (brain, heart, lungs, etc.). So in that sense, I do not think faithful emulation is possible. In fact, I have my doubts about whether we will achieve human comparable AIs before we come to understand more fully the inner workings of the brain, especially in relation to what exists in the zero point field. Perhaps this is a deterministic telological means of reasoning but it is a gut feeling I currently hold and a question worth addressing.
--macrocosm Tue Dec 21 8:38:47 2004
henrik - Tue Dec 21 11:09:38 2004
For the purposes of this thread, let me define a "faithful emulation of the human mind" as being an emulation capable of passing the Turing Test.
--henrik Tue Dec 21 11:09:38 2004
charmed_quark - Sun Dec 26 18:43:21 2004
I don't think that Alan Turing intended the Turing test to be rigorous benchmark of A.I. The paper where he presents it is uncharacteristically sloppy (at one point the test could be interpreted as a computer imitating a man imitating a woman) and it seems to be more a flippant response to his detractors than a serious opinion.
John Searle gave what I consider a sufficient rebuttal to the Turing Test as a measure of AI, with his Chinese Room analogy. While Dr. Searle refuted the usefulness of the Turing Test, I do not agree with him that this alone makes strong AI an impossibility.
Consider the termite. An individual termite is an uncomplicated, single minded critter. Certainly we can believe that a computer could emulate the mind of a termite, if it is not already possible. However as a collective termites are capable of constructing marvelous structures, with internal cooling mechanism and agricultural development, over the course of generations. Now what if a group of 'termite intelligences' were not limited by their own physicality? It is this type of emergent behavior that I believe will develop and be prominent in AI.
I cannot see any real benefit to developing human-like AI other than as a type of technological demonstration. The unpredictable nature of humans makes it difficult to measure the degree of success.
Macrocosm: I'm extremely interested in the physics breakthroughs that you're referring to. Can you provide more information or a link?
--charmed_quark Sun Dec 26 18:43:21 2004
TheLiberal - Sun Sep 30 2007
Computers will be able to completely simulate the inner workings of the human mind by 2030. No special hardware tricks. To brilliant AI programming. Just straightforward emulation, molecule by molecule.
But the human mind is born knowing nothing, so if computer scientists go that route, the the computer will have to learn, just as a human does. There really is no reason to make a comparable Human AI, but we humans seem fascinated with the idea anyway.
--TheLiberal - Se abovedate
