1. What is artificial intelligence?
There are many different kinds of traditional artificial intelligence. Taking the learning agent as an example, he will perceive the environment through sensors, make decisions through the environment, and then evaluate actions and their results in the process of action, so as to learn from them and improve future decisions. Through this experience, he will continue to learn and make better plans and decisions. We can see that it can have a certain perception of the surrounding environment, and with the passage of time, it will make better decisions from it.
Learning agent
For example, the once popular Go AI Alpha Go is like a more advanced self-learning AI. He can observe the chess game, make calculations, make decisions, and continue to learn through the reaction of his opponents. We also saw its amazing record against professional chess players. But does this mean that artificial intelligence can produce autonomous consciousness? This is the next part, specifically understanding what is autonomous consciousness.
2. What is autonomous consciousness?
At the same time, for this problem, we need to realize what is the definition of autonomous consciousness. From the philosophical point of view, self-consciousness is a kind of consciousness acquired after introspection. Professor Zhou Changle of Xiamen University once described the mechanism of people's self-consciousness: "The change of material movement creates everything, and the physiological activities of living things support the neural activities, and the neural activities emerge consciousness (conscious psychological activities). Consciousness feels the physiological performance and guides the realization of intentional psychological activities, thus embodying the principle of knowing everything." Simply put, this is a cycle of "matter-physiological activity-nerve activity-consciousness-mental activity-matter".
But now, for the time being, such a mechanism cannot be achieved. Artificial intelligence is still based on countless machine learning and neural network algorithms. Self-awareness requires a deeper and broader feeling and cognition of the surrounding environment. He can only do what he needs to do according to the algorithm completely, and he can't reach the goal. When people can accomplish all kinds of complicated goals with the help of tools, they will feel tired and excited in the process, and they can understand the purpose of their actions and the actions themselves. This is the lack of mental activity. They can't feel this task except how to complete the accident better according to the algorithm.
summary
Let's go back to the example of Alpha Go. He can only learn to make Go better and better, but he can't change his mood in the process. He won't get more revenge for the next game because he lost, nor will he be happy because he won. He can feel the environment, such as the chess game on the board, but he can't feel the opponent's psychological state, the atmosphere at the scene and so on. From it, we can see that artificial intelligence, as the result of data algorithm accumulation, can only be used to help human beings complete specific content. Artificial intelligence is still limited, it can't have self-awareness, and it will be difficult to achieve it in the future.