Can a Machine Know?
Jun. 13th, 2006 11:55 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
knights_say_nihThis is boring and academic and I need to be able to get at it on a different computer so it's here in the mean time. Don't read unless you like philosophy.
Can a machine know?
This is a question our society has been, and will be dealing with in years to come as our technology advances and Artificial Intelligence becomes more and more of a reality. But taking a step back, we ask ourselves the basic question, ‘can a machine know.’ Not an intelligent, thinking machine but something as simple as a calculator, a laptop. Knowledge, as justified, true belief is not something unique to humans, nor to animals, and it is my belief that within that definition of the word a machine can indeed know. I would argue that machines can know for several reasons.
First, in the similarity of humans to machines, on a fundamental level, and our assumption that we can know. Second, in the way a machine learns, both through studying its surrounding and accepting passed on information as true. A machine has a lack of bias that will never be possible for a human. We have trusted them with tasks, assuming that their knowledge of how to perform them will be correct. Despite the argument that they only learn from passed on information, this is much the same as a human, and they are capable of gathering their own information as well and forming their own conclusions. Though they cannot create, this is not a quality essential for knowing. They can communicate between themselves, and more reliably than humans can. Despite human assumptions and prejudices, and despite the fact that we will never be able to prove this conclusion to any degree of certainty, we must operate under the assumption that machines can know.
Of course, the main problem of knowledge is that we will never be truly sure because we are not, in fact, machines. We cannot know if they are processing what we tell them or merely drawing from a list of set probabilities and finding the logical conclusion. Another problem of knowledge is the human tendency to rebel against anything unknown will colour our perspectives. Furthermore, if it is the case that a machine is only repeating what it has been programmed to see we will never know if it is being programmed to simulate intelligence or has actually evolved to have it. We are not machines and do not have the guarantee that we can trust what they tell us.
Although emotionally the human spirit may rebel against the thought of a machine ‘knowing’ it is not fair to allow this prejudice to colour our perspectives. This is an ethical question we will have to look into in sociological and legal ramifications in the future, as we move closer and closer to developing AI. If we create a sentient entity, will it be fair to give it less than human status because it is made of metal? It may be conscious but will we be able to accept it? It is important to set a precedent of accepting non-humans as conscious.
For that reason alone I would be prepared to say that it is possible for a machine to know.
The idea that knowledge is unique to human beings strikes perilously close to the racist philosophies that expressed the ideas that one sort of person was superior to another for something as superficial as skin colour. The body does not define the knowledge within, and this goes the same for the substance the body is made of. Thomas Franklin said “One day our descendants will think it incredible that we paid so much attention to things like the amount of melanin in our skin or the shape of our eyes or our gender instead of the unique identities of each of us as complex human beings” and if we choose now to limit the definition of knowledge to only include humans, our descendants will be looking at use with the same incredulity.
Multiple religious groups argue that man was created in God’s image, and profess our inability to recreate him any other way as a sign that searching for AI is an unethical pursuit. However, we remain a secular country, and until we revert to a religious based society cannot allow this opinion to colour the laws made by the state.
Knowledge is not something unique to humans. In fact, humans are, in their own way, machines. All human life is carbon based, water based. We are put together the same way, only made out of flesh and blood, a theory first proposed by Renee Descartes in the 17th century. Can we know? Yes, we evidently assume we can. We use our senses to procure information, process and come to conclusions that we ‘know.’ Machines, as mechanical extensions of us, do the same thing. They absorb data from their surroundings, through sensors, or are given it word of mouth by their programmers much in the way our parents teach their children. This does not in any way cheapen this knowledge. In this way, machines can absorb something and then proceed with the assumption that that is the complete truth.
In this way, it’s almost more likely that a machine can know than a human can. Machines can be programmed to do one specific task which they understand inside out and without bias. By being what they are, there is no bias, and thus the belief is more likely to be both justified and true. On top of that, their means of observing, of taking in their environment are more reliable than a humans and more complete, again making their conclusions more likely to be true.
In the past, we have trusted great responsibility to the knowledge of machines. For example, the Automatic Missile Defense System, something designed to launch a nuclear warhead should an attack be sensed. Currently, project PATRIOT is already employed, an automatic section of the USA’s defence program. If this system is put into play, then we will be brokering on the fact that the knowledge of the machines- that nuclear attack is imminent- to change the course of the world by launching an offensive. As a society, we are trained to accept the belief that a machine’s knowledge is absolute, without applying the word ‘knowledge’ to the data in question.
Indeed, the knowledge gathered by machines should be qualified as reliable. Machines do not have the probability for error that humans do, as they are programmed by the best and the brightest of us. In the way the child of especially bright parents will have inherited favorable genetics and been around intelligent conversation at the dinner table. Our knowledge is put into them, and they use it where we could not.
In the novel ‘I, Robot,’ a semi-religious speculative fiction novel by the illustrious Isaac Asimov, the main character questions the machines and their capacity for beauty and creativity. That they well be separated by their inability to create, substituting it instead with reproduction. The question is asked ‘can a robot write a symphony? Can they create something beautiful?’ and the answer is, ‘can you?’ Creation and knowledge are not mutually inclusive. Many humans themselves will never create anything truly unique, original or lasting; however no one questions their ability to know. Knowledge should not be a definition applied immediately to humans and denied to everything else.
Although machines cannot technically communicate, only parrot learned responses to a memorized keyword, they do in fact understand one of the more basic languages completely of their own accord and without the necessary ‘parroting’ implied. Mathematics has often been called the universal language, and a machine can understand and work with it in a way that a human cannot. In this way machines can communicate through the language barrier, thus speaking with each other, or anyone else without risk of meaning being lost in translation.
Perhaps it could be argued that all a machine knows is given to it through humans programming it, and in this way machines are merely a repository for knowledge, however I argue that although this is true, all that a human knows is achieved in much the same way. Although it is easy to sneer and say ‘they know what we give them,’ we, for the most part are much the same. It will be difficult to separate the repetition of facts from true knowledge. Is repeating facts the same thing as knowing? I would say, by the definition provided by our school systems, yes. We know that the world is round but none of us have ever actually been to space and seen that this is true. Passed on knowledge is true knowledge, it is the beginnings of everyone’s education, and will prove as a stage from which machines will work as they evolve into something more conscious.
Can a machine know?
This is a question our society has been, and will be dealing with in years to come as our technology advances and Artificial Intelligence becomes more and more of a reality. But taking a step back, we ask ourselves the basic question, ‘can a machine know.’ Not an intelligent, thinking machine but something as simple as a calculator, a laptop. Knowledge, as justified, true belief is not something unique to humans, nor to animals, and it is my belief that within that definition of the word a machine can indeed know. I would argue that machines can know for several reasons.
First, in the similarity of humans to machines, on a fundamental level, and our assumption that we can know. Second, in the way a machine learns, both through studying its surrounding and accepting passed on information as true. A machine has a lack of bias that will never be possible for a human. We have trusted them with tasks, assuming that their knowledge of how to perform them will be correct. Despite the argument that they only learn from passed on information, this is much the same as a human, and they are capable of gathering their own information as well and forming their own conclusions. Though they cannot create, this is not a quality essential for knowing. They can communicate between themselves, and more reliably than humans can. Despite human assumptions and prejudices, and despite the fact that we will never be able to prove this conclusion to any degree of certainty, we must operate under the assumption that machines can know.
Of course, the main problem of knowledge is that we will never be truly sure because we are not, in fact, machines. We cannot know if they are processing what we tell them or merely drawing from a list of set probabilities and finding the logical conclusion. Another problem of knowledge is the human tendency to rebel against anything unknown will colour our perspectives. Furthermore, if it is the case that a machine is only repeating what it has been programmed to see we will never know if it is being programmed to simulate intelligence or has actually evolved to have it. We are not machines and do not have the guarantee that we can trust what they tell us.
Although emotionally the human spirit may rebel against the thought of a machine ‘knowing’ it is not fair to allow this prejudice to colour our perspectives. This is an ethical question we will have to look into in sociological and legal ramifications in the future, as we move closer and closer to developing AI. If we create a sentient entity, will it be fair to give it less than human status because it is made of metal? It may be conscious but will we be able to accept it? It is important to set a precedent of accepting non-humans as conscious.
For that reason alone I would be prepared to say that it is possible for a machine to know.
The idea that knowledge is unique to human beings strikes perilously close to the racist philosophies that expressed the ideas that one sort of person was superior to another for something as superficial as skin colour. The body does not define the knowledge within, and this goes the same for the substance the body is made of. Thomas Franklin said “One day our descendants will think it incredible that we paid so much attention to things like the amount of melanin in our skin or the shape of our eyes or our gender instead of the unique identities of each of us as complex human beings” and if we choose now to limit the definition of knowledge to only include humans, our descendants will be looking at use with the same incredulity.
Multiple religious groups argue that man was created in God’s image, and profess our inability to recreate him any other way as a sign that searching for AI is an unethical pursuit. However, we remain a secular country, and until we revert to a religious based society cannot allow this opinion to colour the laws made by the state.
Knowledge is not something unique to humans. In fact, humans are, in their own way, machines. All human life is carbon based, water based. We are put together the same way, only made out of flesh and blood, a theory first proposed by Renee Descartes in the 17th century. Can we know? Yes, we evidently assume we can. We use our senses to procure information, process and come to conclusions that we ‘know.’ Machines, as mechanical extensions of us, do the same thing. They absorb data from their surroundings, through sensors, or are given it word of mouth by their programmers much in the way our parents teach their children. This does not in any way cheapen this knowledge. In this way, machines can absorb something and then proceed with the assumption that that is the complete truth.
In this way, it’s almost more likely that a machine can know than a human can. Machines can be programmed to do one specific task which they understand inside out and without bias. By being what they are, there is no bias, and thus the belief is more likely to be both justified and true. On top of that, their means of observing, of taking in their environment are more reliable than a humans and more complete, again making their conclusions more likely to be true.
In the past, we have trusted great responsibility to the knowledge of machines. For example, the Automatic Missile Defense System, something designed to launch a nuclear warhead should an attack be sensed. Currently, project PATRIOT is already employed, an automatic section of the USA’s defence program. If this system is put into play, then we will be brokering on the fact that the knowledge of the machines- that nuclear attack is imminent- to change the course of the world by launching an offensive. As a society, we are trained to accept the belief that a machine’s knowledge is absolute, without applying the word ‘knowledge’ to the data in question.
Indeed, the knowledge gathered by machines should be qualified as reliable. Machines do not have the probability for error that humans do, as they are programmed by the best and the brightest of us. In the way the child of especially bright parents will have inherited favorable genetics and been around intelligent conversation at the dinner table. Our knowledge is put into them, and they use it where we could not.
In the novel ‘I, Robot,’ a semi-religious speculative fiction novel by the illustrious Isaac Asimov, the main character questions the machines and their capacity for beauty and creativity. That they well be separated by their inability to create, substituting it instead with reproduction. The question is asked ‘can a robot write a symphony? Can they create something beautiful?’ and the answer is, ‘can you?’ Creation and knowledge are not mutually inclusive. Many humans themselves will never create anything truly unique, original or lasting; however no one questions their ability to know. Knowledge should not be a definition applied immediately to humans and denied to everything else.
Although machines cannot technically communicate, only parrot learned responses to a memorized keyword, they do in fact understand one of the more basic languages completely of their own accord and without the necessary ‘parroting’ implied. Mathematics has often been called the universal language, and a machine can understand and work with it in a way that a human cannot. In this way machines can communicate through the language barrier, thus speaking with each other, or anyone else without risk of meaning being lost in translation.
Perhaps it could be argued that all a machine knows is given to it through humans programming it, and in this way machines are merely a repository for knowledge, however I argue that although this is true, all that a human knows is achieved in much the same way. Although it is easy to sneer and say ‘they know what we give them,’ we, for the most part are much the same. It will be difficult to separate the repetition of facts from true knowledge. Is repeating facts the same thing as knowing? I would say, by the definition provided by our school systems, yes. We know that the world is round but none of us have ever actually been to space and seen that this is true. Passed on knowledge is true knowledge, it is the beginnings of everyone’s education, and will prove as a stage from which machines will work as they evolve into something more conscious.
