Re: Can a notebook in the cell suffer ...
This is a response to the post http://habrahabr.ru/blog/artificial_intelligence/46316.html , which for some reason did not fit in the comments.
The logic of the author’s “reasoning” differs little from the “logic” that made the ancients believe that the sun revolves around the earth. Appeal to subjective emotions (“well, how can iron feel?”) It is impossible to strictly prove objective phenomena, let alone talk about attempts to prove or disprove someone else’s subjectivity. With such reasoning, it is possible to get to the clarification of the type “Do Protestants have a soul and can they be burned at the stake?”. In addition, an experiment with a notebook or “Chinese room” implies the presence of an agent, respectively, and the reason in this case (“knowledge of Chinese”) is possessed not by a stupid agent, and not by a notebook by itself, but by a system of both of them. A system is always more than the sum of its components!
')
There is such a thing as the invariance of information about the carrier. If there is a system consisting of anything — neurons, wires or notebooks in a cell — that behaves like a human brain, learns, communicates, etc., demonstrates reactions similar to those observed in humans, then this actually means that of this structure, the object is aware of itself as a rational being, and it is desirable to recognize it as such from moral and ethical considerations, for reinsurance.
In the above example, the behavior of the brain is emulated at the lowest level, starting with neurons. High-level structures (feelings, thoughts) are encoded in low-level like TCP to IP, HTTP to TCP, and so on. From the point of view of suitability for transferring information, notebooks with cells are no different from fibers with photons, except for potential speeds. As well as from neurons. To build such a brain emulator, it is necessary to “climb” into the human brain and reproduce its low-level device.
On the other hand, it is theoretically possible to construct an AI using the “black box” method, not getting inside the low-level structure, but imitating its external behavior. Suppose the developer does not know exactly how the low-level structure works, but in trying to reproduce its behavior, he is developing a structure that is somewhat similar to the original one. So an interesting question arises: is this mind realized on the basis of feelings and sensations similar to ours - or some other?
And is it possible to determine unequivocally whether these xenochuvies are worthy of respect, or can they be treated as you please?
The approach demonstrated by the author - a subjective assessment - has been used in history for a long time and led to questions like "Do Protestants have a soul and can they be burned at the stake?" Modern society as a whole is already ripe to recognize the rights of any born person, as well as some animal rights (for the time being it is a tribute to the emotions of their advocates, and not an understanding of real needs). Today, no civilized person will ever wonder whether Mr. or Mrs. A. has a soul. The battlefield today is the rights of a human being before birth, great apes, potential visitors from space and our potential offspring - intellectual robots and so-called E- creatures - to start on a decent life.
In general, the very idea of ​​the independence of the information process (“soul”) from a specific given carrier (“body”) was expressed in the first phrase of the Bible: “In the beginning was the WORD” (in Greek, “logos”). The WORD can be expressed as anything and on anything, even though the cells in the notebook, from this it does not cease to be so. The criterion proposed by the author is a kind of chauvinism (preference of one type of media over the others), and the error of the author’s reasoning is in an attempt to pass the relative to the absolute, for example, “real time” —which is real? The processor is the same notebook with very quickly filled and erased many cells. And the brain - a few gigattradok. It is important not WHAT, but HOW.
If you look at the history of human civilization, the question of the relationship of the subject to another entity (man, animal, alien, robot, E-essence) was decided at the very beginning by instincts (“interests” of the organism), then by interests of the individual, group, society and all of humanity , the next step is the Cosmos, in which there can be Solaris and whatever. Any particular solutions with a hard division into black (“trembling creature”) and white “I have the right”) may turn out to be inadequate. Controversial cases should be very carefully studied (in fact, not in form) and only then can one make any definite conclusions about the admissibility or inadmissibility of certain operations with a specific instance of AI. There are no such instances of AI at the moment, but a part of the rat's brain has been emulated with a speed of 10 times slower, recreating working structures of the highest level is a matter of time, and it is better to prepare for this in advance so as not to get a surprise in the form of an unfriendly AI or hidden genocide, with all its consequences.
The whole question of philosophy (humanism, ecohumanism, transhumanism, etc.) and science (AI, bioinformatics, neurobiology, cognitive science, synergetics, ...) deal with the question that the author tried to answer. even when the mechanisms defining the human in man are studied and defined with a margin of safety, many border issues will arise, for even if the settlement of relations within one biological species has only recently been somehow solved at the top level (UN, HRH), and more In some places these issues are still super-actual in some places, let alone other forms of the existence of the mind. The main thing is not to try to resolve this issue categorically “once and for all” (“to burn the Protestants, not to burn the Catholics”), but to direct the resolution process to a safe course.
As for the Turing test, it rather leaves a safety margin, although it has limited scope (only for human-like high-level intelligence; the traditional Turing test is not applicable to higher animals, babies or downs), but the whole idea is clear. As the Americans say, if something behaves like a duck, quacks like a duck, swims like a duck, cleans feathers like a duck, then most likely this is a duck.
The logic of the author’s “reasoning” differs little from the “logic” that made the ancients believe that the sun revolves around the earth. Appeal to subjective emotions (“well, how can iron feel?”) It is impossible to strictly prove objective phenomena, let alone talk about attempts to prove or disprove someone else’s subjectivity. With such reasoning, it is possible to get to the clarification of the type “Do Protestants have a soul and can they be burned at the stake?”. In addition, an experiment with a notebook or “Chinese room” implies the presence of an agent, respectively, and the reason in this case (“knowledge of Chinese”) is possessed not by a stupid agent, and not by a notebook by itself, but by a system of both of them. A system is always more than the sum of its components!
')
There is such a thing as the invariance of information about the carrier. If there is a system consisting of anything — neurons, wires or notebooks in a cell — that behaves like a human brain, learns, communicates, etc., demonstrates reactions similar to those observed in humans, then this actually means that of this structure, the object is aware of itself as a rational being, and it is desirable to recognize it as such from moral and ethical considerations, for reinsurance.
In the above example, the behavior of the brain is emulated at the lowest level, starting with neurons. High-level structures (feelings, thoughts) are encoded in low-level like TCP to IP, HTTP to TCP, and so on. From the point of view of suitability for transferring information, notebooks with cells are no different from fibers with photons, except for potential speeds. As well as from neurons. To build such a brain emulator, it is necessary to “climb” into the human brain and reproduce its low-level device.
On the other hand, it is theoretically possible to construct an AI using the “black box” method, not getting inside the low-level structure, but imitating its external behavior. Suppose the developer does not know exactly how the low-level structure works, but in trying to reproduce its behavior, he is developing a structure that is somewhat similar to the original one. So an interesting question arises: is this mind realized on the basis of feelings and sensations similar to ours - or some other?
And is it possible to determine unequivocally whether these xenochuvies are worthy of respect, or can they be treated as you please?
The approach demonstrated by the author - a subjective assessment - has been used in history for a long time and led to questions like "Do Protestants have a soul and can they be burned at the stake?" Modern society as a whole is already ripe to recognize the rights of any born person, as well as some animal rights (for the time being it is a tribute to the emotions of their advocates, and not an understanding of real needs). Today, no civilized person will ever wonder whether Mr. or Mrs. A. has a soul. The battlefield today is the rights of a human being before birth, great apes, potential visitors from space and our potential offspring - intellectual robots and so-called E- creatures - to start on a decent life.
In general, the very idea of ​​the independence of the information process (“soul”) from a specific given carrier (“body”) was expressed in the first phrase of the Bible: “In the beginning was the WORD” (in Greek, “logos”). The WORD can be expressed as anything and on anything, even though the cells in the notebook, from this it does not cease to be so. The criterion proposed by the author is a kind of chauvinism (preference of one type of media over the others), and the error of the author’s reasoning is in an attempt to pass the relative to the absolute, for example, “real time” —which is real? The processor is the same notebook with very quickly filled and erased many cells. And the brain - a few gigattradok. It is important not WHAT, but HOW.
If you look at the history of human civilization, the question of the relationship of the subject to another entity (man, animal, alien, robot, E-essence) was decided at the very beginning by instincts (“interests” of the organism), then by interests of the individual, group, society and all of humanity , the next step is the Cosmos, in which there can be Solaris and whatever. Any particular solutions with a hard division into black (“trembling creature”) and white “I have the right”) may turn out to be inadequate. Controversial cases should be very carefully studied (in fact, not in form) and only then can one make any definite conclusions about the admissibility or inadmissibility of certain operations with a specific instance of AI. There are no such instances of AI at the moment, but a part of the rat's brain has been emulated with a speed of 10 times slower, recreating working structures of the highest level is a matter of time, and it is better to prepare for this in advance so as not to get a surprise in the form of an unfriendly AI or hidden genocide, with all its consequences.
The whole question of philosophy (humanism, ecohumanism, transhumanism, etc.) and science (AI, bioinformatics, neurobiology, cognitive science, synergetics, ...) deal with the question that the author tried to answer. even when the mechanisms defining the human in man are studied and defined with a margin of safety, many border issues will arise, for even if the settlement of relations within one biological species has only recently been somehow solved at the top level (UN, HRH), and more In some places these issues are still super-actual in some places, let alone other forms of the existence of the mind. The main thing is not to try to resolve this issue categorically “once and for all” (“to burn the Protestants, not to burn the Catholics”), but to direct the resolution process to a safe course.
As for the Turing test, it rather leaves a safety margin, although it has limited scope (only for human-like high-level intelligence; the traditional Turing test is not applicable to higher animals, babies or downs), but the whole idea is clear. As the Americans say, if something behaves like a duck, quacks like a duck, swims like a duck, cleans feathers like a duck, then most likely this is a duck.
Source: https://habr.com/ru/post/28920/
All Articles