I do not think that there are some important "philosophical" reasons why it would be impossible to build a machine endowed with intelligence, including the ability to use natural language in the same way as we use it. You may even think that the machines performing simple algorithms can have an awareness that the implementation of simple instructions may be to them a simple "survival." Yet I believe that consciousness is a natural feature of simple physical processes. Similarly, I believe that the human brain is a biological computer - an extremely complicated, capricious, unpredictable, computer, however, not a "controlled intangible spirit of the relay signals." Hindu sages say that all is Brahman. If they were right, that would mean that, and the computer is Brahman. In what would it be worse? Conscious and intelligent computer may even possess free will, if the concept is understood as a feeling of compatibility of the decision of the value system is adopted. It can be humanly capricious and unpredictable. Theoretically, "may", however, between theory and practice gap opens sometimes impassable. Technical and theoretical problems that are facing the issue of artificial intelligence and the task of verification of hypotheses related to the "artificial" consciousness is in my view, enormous. On the problems of these, but some possibilities and I want to write some comments in today's entry.
Firstly, I think it is very important to make a distinction between intelligence and consciousness (or "awake", the ability to feel, to be subject of survival). Something like a cross between these two concepts to be aware of "higher order" linked to the rich associational structure, including - but not necessarily and not only - of language, semantics, and thus the backbone to express any thoughts, as well as aesthetics, ethics, etc. .. With it connects to our artistic sensitivity, sensitivity to the beauty of nature, to poetry, the beauty of mathematical formulas, the beauty of simplicity of proofs in mathematics and logic. This phenomenon of the human mind causes much confusion and is sometimes confused both with knowledge and with intelligence, meanwhile, although there is because of our intelligence and sensitivity, thanks to the richness of the structure of the human soul, he is content to "experience" rather than an actual experience. An example of confusion about the concepts of consciousness, intelligence, and "understanding" is a discussion around the famous "experiment devised" by John Searle (see "Chinese Room" ), where - in my opinion - everyone understood the concept differently polemicists, for which he had argue.
Some confuse it with the awareness of intelligence. They believe that conscious beings can only be equipped with intelligence and intelligent information processing that is always accompanied by consciousness. That need not be so, the easiest way we shall see when we notice that in a very intense and conscious way, we can live united in no way engaging our intelligence - physical pain, sexual arousal, 'animal' fear of appearing in a time of danger. If, in turn, we consider the enormity of complex calculations, which are carried out - for example - the brain of the host to provide a football striker in the penalty opponent and believe, that his whole attention is then focused on the movements of the goalkeeper, on the movements of the advancing defenders and the same shall be receiving the ball instinctively and out of consciousness, then we understand that the processing of information and awareness, not necessarily the same thing.
For example, the player to go back even further, presenting his own, very simple, but extremely technically difficult to verify the hypothesis concerning the nature of consciousness.
But let's start from the intelligentsia. In the 40 pioneer in computer technology, a prominent English mathematician and kryptoanalityk - Alan Turing proposed test that could show whether the machine, or rather performed by that program is or is not intelligent. This test would consist in placing in two separated areas: the computer has undergone a test in one man and moves in the second test. The tester does not know whether, in the second room is a man or a computer and using klawiaturowo - monitors "conversations" have it resolved. If he fails to recognize that, in the second room has a computer, then we say that the Turing test has been passed. This "behavioral" test seems to be methodologically correct. There is probably no other way to verify someone's intelligence (in this case is meant, as a capacity for reasoning in a human manner, in a manner appropriate intelligent people), and an interview with him and to examine the reasonableness of his speech. Although I do not deny nor the capability to build intelligent machines or the fact that about someone (something) intelligence to know we can only study it from "outside", I believe, however, that to the Turing Test to be considered credible, must be fulfilled several additional conditions, most important of which is one hundred percent effective during the "exam". If anyone manages to "break" mindset algorithm, then the person will be able to betray his "way" of other testers, and since then the machine does not pass the test any longer. Meanwhile, we recognize that intelligence is a trait that we can not lose, only because of something someone told on us. We understand it as a feature of "internal". For exactly the same reasons, the description of the test algorithm should be widely available. This stipulation clearly raises the bar for all aspiring to pass the Turing Test automata. Keep in mind that the intelligence of each algorithm should be constantly treated as a hypothetical. As the algorithm will take the next test, our belief in "truth" of his intelligence would of course increase, but if someone ever manage to "catch" it to the evident "inhuman stupidity", then it turns out that he was not smart never (! ), and not only when it is discovered. Scientific hypotheses and theories, including those which deal with the intelligence of specific machines, we can only forge. We can never carry their proof - unless we describe the nature of human intelligence and prove mathematical theorems in the form - but I think this idea as crazy. Probably
intelligent machine will have to have human connotations, and it is unknown whether this can be achieved not by taking part in the human good and bad - in the suffering, humiliation, pain, weakness, ułomnościach, shame, pride, ambition, triumph, pleasure, joy, satisfaction , surprise, a variation of what life brings. Is the humor, irony, malice, or betray our humanity mistakes, slips of the tongue, errors in general are possible without a lot of wealth born of the experience?
now come to consciousness. I believe, or at least close to this I am to believe that consciousness is accompanied by all the physical processes taking place in the universe, that it is somehow a fundamental property of existence. And that is why we speak of artificial consciousness is the same misconception as to talk about "artificial electron." It may seem that he deny what I wrote a few paragraphs above about piłkarzu "unconsciously," the host application. Contradiction however, disappears when we notice that in order consciousness could "demonstrate" that could "testify about myself," must be accompanied by memorizing survival. Experience, which has not been saved anywhere immediately "dies" and verify that it occurred that it was "conscious", it becomes impossible. When our football, even as he learned to master the avatar given away the ball, these exercises focus on all of his attention. Then, each receiving the ball had been aware of. Once this technique has been assimilated, the player did not have longer focus on the parties. He did not need them anymore, "remember" and so they started to disappear from his consciousness. Notice that the word "his" there is no sense only in the context of a specific human memory. That memory is the only glue "I", the only factor that separates the self from the rest of existence. "I", this (or that), who remembers my experiences, and who (or what) can remember them.
My hypothesis assumes that any algorithm capable of storing any simple event, and also for reference to what is remembered is at the time of such a process in a simple way conscious. This hypothesis seems difficult to challenge on the grounds of logic alone, and perhaps sometime in the distant future, we will be able to falsify it, or made more credible by some experiment. Verifying this hypothesis, however, will connect more with carrying out experiments on humans, not machines, because people just are beyond doubt aware. If ever our knowledge about the brain and technology have gone far enough that we can "protein" survival memorize in silicon and silicon experience to remember in the "protein", perhaps uprawdopodobnimy hypothesis that the silicon is capable of "inner experience" and that the content of consciousness may be something very simple, separate from the context of human experience.
Some fear that the emergence of intelligent and conscious machines will entail enormous, difficult to predict ethical problems. I do not agree with that. Even if the machine will create a truly and humanly intelligent, conscious, which has even some digital feelings her "personality" will almost certainly very different kind than ours. Easier for me to imagine that we will in some way "more than" than it would be "human" without participation in experiences such as humiliation after losing a fight, the joy of winning the match and the smell of their sweat, as awareness of the coming of old age and fear of death . While we can succeed in understanding it, will never feel the same way as we do. Creation of "standing in the server" machine with a higher intelligence than our own will almost certainly be technically easier than building an android which we give to human fate. With this machine, it will not unite us no biological relationship: it will be the "other species" and a million times more "different" than any other animal species living on Earth.
Although it is "theoretically possible", is "artificial man" ever actually built?
In the seventies, after a decade of success in the conquest of space around the Earth, after a few landings on the moon man, it seemed that the era of tackling more areas of the universe, flight to the stars, or at least mass manned spaceflight vessels for further planets of the Solar System will inevitably come - if not in the twentieth century, at least in the early twenty-first century. We already know that this was not true. Not true, not only because it became too difficult technically, but also because it turned out to be simply unaffordable. Instead of flights to Mars but we have satellite television, personal computers in almost every home, internet, mobile phones. Technology development went in the direction of providing what is perhaps less glamorous, but more useful and easier to achieve. Why should we seek to force the bestowal of computers, human intelligence, since their "inhuman" data processing speed can be useful to us much more if we use it in another way? Today celebrated successes of artificial intelligence in any way, do not take the machine to imitate the way human reasoning. And the path goes rather further development of information technology, and project sponsors picking itself the goal of full "humanization" of the machine will find smarter ways to invest money.
0 comments:
Post a Comment