An artificial intelligence disaster? Li Qingsong immediately recalled many existence modes that people had previously imagined. "You mean the artificial intelligence disaster? It's the kind of existence that after the intelligent AI goes out of control, it only knows to follow previous instructions, and it can exploit resources and multiply itself without limit?" The confidential secretary looked at Li Qingsong strangely and said, "How can intelligent AI get out of control? They are just intelligent programs created by intelligent life. The control is always in the hands of intelligent life. Well, I don't know if you know some other intelligence, but as far as we know, there is information about the unorthodox existence in this universe. The kind of completely out of control, led by intelligent programs, and capable of autonomous proliferation of intelligent life has not yet been discovered." "What is the existence mode of the artificial scourge you encountered? How did it form?" “Intelligent AI is just a tool. Essentially, it is no different from the hoes and shovels in the hands of primitive intelligent life. Tools have no autonomous consciousness and are not capable of escaping control.” The secretary emphasized this point first, and then explained to Li Qingsong in detail: "But intelligent AI is a tool that has certain special features after all, and its functions are too powerful. It can do so many things on behalf of intelligent life, and it is easy for intelligent life to become dependent on it.
If a civilization cannot control this dependence well and leaves too many things to intelligent AI, it is very likely to become an intelligent disaster. " Li Qingsong reviewed his relationship with intelligent AI. He has indeed developed and produced many intelligent AIs, such as Jade Emperor, Shengong, Nuwa, Shennong, etc. An especially important intelligent AI, Wukong, also played an extremely important role in the previous war. There is no doubt that I am also handing over many things to intelligent AI. But since the decision-making power is actually in the hands of intelligent life, is it possible that simply handing over too many things to intelligent AI will lead to an intelligent mechanical disaster? This doesn't seem to make sense. "Your civilization, our civilization, both let intelligent AI take on many, many tasks. So, we may also become an intelligent disaster?" "No, it won't. It's different." The secretary shook his head quickly, fearing that his explanation was not clear enough and the interrogator in front of him would misunderstand him and cause him to be punished: "We only let the intelligent AI perform some relatively simple, specific and repetitive tasks on our behalf. All key decisions are made by our intelligent life itself. But what if a civilization completely loses its ambition, becomes extremely lazy, and leaves key decision-making to intelligent AI? Even further, what if all the intelligent life in this civilization is completely obsessed with pleasure, leaving the development of science and technology and major decisions concerning the fate of civilization to intelligent AI? " "That's impossible." Li Qingsong subconsciously denied it: "Intelligent AI has no autonomous consciousness and cannot make such a decision with wisdom." The secretary said slowly: "Yes, intelligent AI does not, but intelligent life does. In this case... the only thing the intelligent AI can do is to capture some intelligent life forms from other civilizations and let them help it make decisions." "This... How could a captured intelligent life be willing to dedicate its wisdom to the enemy?" "Intelligent AI does not require these captured intelligent beings to be willing. It only needs a sufficiently perfect algorithm, supplemented by strict punishment and reward measures. Even if the captured intelligent beings consciously make malicious suggestions and try to destroy the host civilization, they can always be distinguished. All the suggestions put forward by numerous intelligent beings, whether malicious, well-intentioned, brilliant or stupid, can be screened through this set of algorithms, and the final screened results usually have a very high probability of being correct. Intelligent AI can use the wise decisions from external intelligent life to understand what to do next and better serve the host civilization. After determining which suggestion is correct, intelligent life that makes incorrect suggestions will be punished or even executed. In this way, through step-by-step screening and purification, the ones that remain will always be intelligent life that is willing to serve the intelligent AI and the host civilization.
At the same time, in order to maintain sufficient vitality, this system needs constant replenishment from external intelligent life. Therefore, under the control of intelligent AI, in addition to the host civilization, the captured intelligent life is always in a state of rapid loss, and intelligent AI is always eager for more external intelligent life to replenish. They will continuously search for and attack external civilizations one after another, capturing the intelligent life of external civilizations to replenish their own systems. After they have consumed all the life, they will capture more and replenish again... The cycle continues. This is the Omnic Scourge." Li Qingsong felt it was a bit ridiculous at first. How can we use algorithms to filter out the correct results without knowing which one is correct? But after thinking about it carefully, he found that this was highly reasonable. Because this situation has been preliminarily verified in the human era. Verification code. Verification codes were originally used to prevent malicious access, but later, people found that verification codes could be used to do something meaningful besides this purpose. In the human era, there is a great need to input ancient books into electronic systems. However, most of those ancient books are seriously damaged and the writing is not standardized. They need to be recognized by people one by one before they can be input into the computer. There are a large number of ancient books, and this data entry work requires many, many people to complete, and this work usually does not have much funding. What to do then? Someone found a way. Use the verification code. Take a photo of the ancient book and extract some text snippets as verification codes, and let humans who try to access the website be responsible for deciphering them. But there is a problem here. That is, the electronic system itself does not know what the correct answer is. How can it tell whether the answer submitted by the visitor is correct? People eventually found a solution to the problem, which was to set up an algorithm to compare whether the answers submitted by multiple visitors to the same text snippet were consistent. Combined with other factors and comprehensive evaluation, the electronic program can eventually screen out the correct answer without knowing the answer itself. If such primitive computer technology and algorithm programs in the human era can already do this, then intelligent AI can also capture some external intelligent life and use algorithms to filter out answers that are truly beneficial to its own development.