"The main target of artificial intelligence international mainstream academics is weak artificial intelligence, and few people are committed to strong artificial intelligence. So, is this because strong artificial intelligence is 'too difficult', so everyone's 'returning to the next'? Otherwise. Most of the artificial intelligence researchers believe that they can't do or should not do it." Some time ago, Zhou Zhihua, a professor at Nanjing University and executive deputy director of the State Key Laboratory of Computer Software New Technology, made an article with a clear view - serious scholars You should not touch artificial intelligence.
This seems to have poured cold water on the development of artificial intelligence, but it can also ease the anxiety of Hawking and Musk. The threat they fear is actually a threat from strong artificial intelligence. If artificial intelligence is limited to weak artificial intelligence, it will only be a clever and obedient assistant.
So, is strong artificial intelligence really a human Pandora's Box?
Research consciousness, the unbearable weight of artificial intelligence
Zhou Zhihua pointed out that the so-called strong artificial intelligence is an artificial object that meets or exceeds the level of human intelligence. It has the mind and consciousness, can act according to its own intentions, and can also be regarded as "artificial intelligence."
At present, the progress and success of artificial intelligence are concentrated on "weak artificial intelligence." Auto-driving, chess, machine vision, expert systems, etc., which people talk about, have nothing to do with strong artificial intelligence. Zhou Zhihua believes that there is no need to have a relationship - if people's goal is to create "tools", then considering a specific type of intelligent behavior is enough, why bother to consider independence?
Moreover, in terms of implementation difficulty, there is no way to consider it.
"To achieve strong artificial intelligence, we must first understand how human intelligence is going. But studying the essence of human intelligence is not the main task of the artificial intelligence discipline." Qin Zengchang, an associate professor at the School of Automation, Beijing University of Aeronautics and Astronautics, told the Science and Technology Daily reporter Consciousness and emotion are an ultimate scientific issue. It is indeed very attractive, but at present the artificial intelligence community may not be able to afford the task of solving this problem.
Of course, many researchers in related disciplines such as biology and neuroscience are looking for ways to uncover the mysteries of the brain. In Qin Zengchang's view, the path of strong artificial intelligence is extremely long, and it may be necessary to start from simulating the brains of insects, fish and mammals, and then step by step to simulate the human brain.
"Strong artificial intelligence is still too far." Qin Zengchang said, "Don't say that we don't know much about nerves and brains. Even if we fully understand it thoroughly, we may not be able to copy strong artificial intelligence."
Yan Zhenming, director of the Interpersonal Interconnect Laboratory at Sun Yat-sen University, believes that consciousness is not what you want, you can have it.
“Any artificial system that does not rely on materials that already have conscious functions, unless there is sufficient reason to conclude that the mechanism or content that was introduced during the artificial process and that is followed by consciousness, we must consider the system to be like the original matrix. The material is not as conscious as it is, no matter how close its behavior is to the behavior of the subject." This is the "artificial intelligence reversal Turing criterion" proposed by Zhen Zhenming. He believes that all artificial machines will not have real consciousness before they enter quantum mechanics.
Study strong artificial intelligence, may wish to develop specifications in advance
The road ahead is really difficult, but is it really impossible?
To continue to discuss this issue, we must return to the definition of strong artificial intelligence. In fact, the academic community does not have a unified view of what is strong artificial intelligence.
Liu Wei, director of the Human-Computer Interaction and Cognitive Engineering Laboratory at Beijing University of Posts and Telecommunications, believes that the intelligence of human-computer integration is strong artificial intelligence, and it will inevitably be realized in the future.
Human-machine integration is the synergy between the intelligence of people and the intelligence of machines. People have knowledge, machines are longer than collecting data; people have experience and common sense, machines are longer than axioms; people have intuition, and machines are longer than logic. When people and machines have enough tacit understanding, people can understand how machines view the world, and machines can be familiar with people's thoughts. Future machines can also have some special specific intentions (to make up for the lack of human cognition). When the two become partners or even confidants, strong artificial intelligence will come.
Zhou Zhihua believes that the creation of strong artificial intelligence has an independent consciousness. It may not be willing to serve human beings. If strong artificial intelligence emerges, human beings will face a huge crisis of survival. Liu Wei pointed out that if strong artificial intelligence is the intelligence of human-computer integration, then the decision-making will always be human, which can skillfully solve public opinion and avoid possible "alternative crisis."
"This is not something that can't be done, but I feel that I need to be very cautious. That is to say, when we are close to the mystery of revealing consciousness and intelligence, we really need to be cautious about every step of the next development." Qin Zengchang stressed .
So, does the "box" of strong artificial intelligence need to be kept on hold?
"Speaking not to study strong artificial intelligence, this is somewhat wishful thinking. Because strong artificial intelligence is an inevitable trend of scientific development." Liu Wei believes that after the emergence of strong artificial intelligence, as far as being taught or teaching is bad, it depends on human beings. The ability. “It’s better not to ask the relevant departments to start thinking about the ethical issues that may be faced in the future, and to introduce corresponding guidelines to minimize possible harm.â€
Shenzhen Happybate Trading Co.,LTD , https://www.happybateprojector.com