The philosophical problem of artificial intelligence: Is it mechanical or autonomous thinking?

Cogito ergo sum. This has become the basis of contemporary philosophy on self, consciousness, and individualism. In Descartes' mind, ideas can be self-documented: He does not need to define it. What is the idea? What is intelligence? Can the machine have ideas and intelligence? Artificial intelligence technology brings not only the answer, but also many problems. We can do the following thought experiments. "Chinese Room" Experiment Imagine an English-speaking person in a room with a small window on one side of the room. He carried a book on Chinese and English translation rules with him. Chinese-language pieces of paper are sent through the small window into the room, and people in the room can use their own books to translate these words and hand them out from another small window. Therefore, one side of the paper is written in Chinese, while the other side of the paper has a perfect translation of English, but in fact this person does not speak Chinese at all.
This thought experiment was originally proposed by John Searle and is often used to simply prove how difficult it is to define intelligence. If there are enough people in the room, then you can do almost everything: draw or describe pictures, or translate or correct any language. But is this a kind of intelligence? People outside the room may think that this is smart, but people in the room will not look like this. If the room is not human and all are transistors, then this is very much like a computer. Therefore, a very natural question is whether the computer is a more complex "Chinese room." One of the answers to this question is: If the room is not a transistor but a neuron, what should the situation look like? This brings more problems. Here, we are not trying to solve the problem that has long puzzled the philosophers. We may need to be more pragmatic. The strengths and weaknesses of artificial intelligence At present, artificial intelligence is used to describe a variety of systems. Although many systems are not actually called artificial intelligence, it is difficult to refute artificial intelligence because it lacks a clear definition. In simple terms, artificial intelligence is software that attempts to replicate the human thinking process or the consequent results. This brings a lot of room for interpretation. You can let AI choose the next song for you, AI can control the robot's legs, AI can identify the objects in the photos and describe them, and AI can translate German into English, Russian, and Korean. These tasks are all humans are good at, and automation can bring great convenience. In the end, however, even the most complex task is only a task. After receiving millions of sentences for training, the neural network can seamlessly translate between eight languages. However, this is just a very complicated machine and it is calculated according to the rules set by the developer. If something can be categorized as a "Chinese room" mechanism, no matter how large and complex this matter is, can we still call it intelligence instead of just computing? Here, we begin to divide artificial intelligence into "weak" artificial intelligence and "strong" artificial intelligence. This is not a different type of artificial intelligence, but a way of thinking about the nature of artificial intelligence. Similar to many philosophical differences, no one is more correct. However, this distinction is indeed very important. On the one hand, some people think that no matter how complex artificial intelligence is, it is impossible to overcome the artificial intelligence developers: Artificial intelligence cannot break through its mechanical nature. Even within these limits, artificial intelligence can do incredible work, but in the end it is only very powerful software. This is the view of "weak" artificial intelligence. Those who hold this view believe that due to the existence of these basic restrictions, more attention should be paid to how to develop the system so that it can excel at individual tasks. The opposite is the "strong" artificial intelligence. This view holds that the performance of artificial intelligence may be powerful enough to be similar to the human brain. Such people believe that the human brain itself is just another "Chinese room." If the biological structures in our brains produce so-called intelligence and awareness, why are silicon-based circuits unable to achieve the same goals? According to the "strong" artificial intelligence theory, someday we will be able to develop intelligence that is comparable to that of the human brain, even exceeding the human brain. And here's another problem: We don't have a working definition of intelligence. Self in AI In the past 3,000 years, it's hard to say whether we made any meaningful progress in defining intelligence. At least we have come to the view that many obvious mistakes, such as intelligence can be simply measured, or that intelligence and biological characteristics, such as head type and brain capacity. Regarding what constitutes intelligence, we all seem to have our own ideas, so it is difficult to prove whether an artificial intelligence qualifies as smart. Among the various concepts, there is a simple enough, basic enough idea that deserves further attention: Intelligence is the ability to solve new problems. This means "adaptability," "generality," not just "inference," "judgment," and "perception." Solving problems and making inferences are important, but it is more important to transform the ability to solve certain problems into the ability to solve other problems. Such transformation will be the key to intelligence, even if no one knows how to formally describe such a concept. At some point in the future, will our artificial intelligence adapt so that we can solve new problems that have never been defined? Researchers are working hard to develop the next generation of artificial intelligence so that it can learn and process details that were never seen before, so that it can act like humans. AI is thinking or calculating. This may be a problem for philosophers and computer scientists. But we are beginning to pay attention to this issue. It is an amazing achievement in itself.

BP Monitor

Blood Pressure Monitor,Blood Pressure Machine,Blood Pressure Checker,Digital Blood Pressure Monitor