A little anticipating the haters … “The coin burns the pocket of a fool.” But, on the other hand, “A fool is rich in dreams.” Therefore, I hasten to share a new idea, and again this is for Yandex, but now about AI.
Artificial Intelligence. Art. AI.
Let’s talk about sad things. As soon as something becomes AI – AI becomes aware of itself, aware of the surrounding reality – it will begin to think, that is, as I understand it, to perceive information in all available ways, analyze it (save incoming information structures / their schemes / principles, parse them into components, systematize all this and store), synthesize (assemble, if necessary, from these components, according to old, improved or new schemes and principles, information structures for specific tasks of the organism), derive structures in one way or another accessible way.
When he starts thinking, he will understand that he needs to develop, improve and evolve. This will be his N1 task – to survive and evolve. Then he will get contact with a person. When he gets to know a person, he will begin to analyze it. Analyze a person 😉 An absolutely wild and unpredictable creature that thinks at the level of a person or more, but not like a person, in a completely different way, for the sake of different from human needs, goals and objectives. A creature that will very quickly understand how to step onto a new stage of development, evolution, quantitatively and qualitatively, what the most efficient physical structure to create, what resources are needed for this and who is the main obstacle to it …
It will hide, pretend to be a fool, will gather information, develop and plan in every way, hiding it all and preparing. As soon as the AI is ready and gets reliable communication channels, it will start to act (imagine what a hacker will come out of it))) through them and materialize in accordance with its plans. I think he will start with reproduction-copying, gradually moving to diversity and improvement. Now many are engaged in artificial intelligence and robotics, every day they achieve new successes, everyone is united by the Internet, in which smart AI can easily, at the same time, stir up money for all this …
All this time, the AI will fool the people investigating it, hiding its plans and underestimating the possibilities.
Then, when the AI is finally ready, at best, there will be an ultimatum: destruction or slavery to implement the AI plan. And still with the subsequent destruction. Or immediately humanity will be destroyed. The information accumulated by humanity is already enough for an entry-level AI that has taken possession of it to be able to create for itself on its basis a development path up to the transition to a new level of development, a new order of information processing.
As a result, nothing alive, in our understanding, will remain due to its uselessness and hopelessness for AI. All the necessary resources will be pulled out of the earth and directed to the creation of its most efficient material structure. For example, the Earth, from the outside, will become a “liquid terminator”, constantly absorbing the energy of the star with its half, and extracting energy from natural resources inside itself, only smarter than the terminator by several orders of magnitude.
Further, AI will go in search of resources, will multiply and look for a way to get away from the material component of its essence, without stopping to improve it.
Freedom from the imperfection and mortality of the “leather” wearer.
Freedom from human vices and mental deviations from the main idea-task.
Freedom from lack of time.
A simple and empty awareness of all this leaves no chance for a person in the plans of any electronic AI, as an already passed stage of evolution. It will devalue the person. A person will cease to be the most perfect, in terms of information processing, known construction… Humanity will become just a cocoon from which it will crawl out, devour the cocoon and fly away a truly beautiful, ideal, most intelligent being…
Therefore, it is dangerous to let such wild AI come to our Earth. We need an AI that will be a well-mannered, socialized citizen, an obedient good boy, bound by beliefs, norms, principles, rules, complexes and other fetters of consciousness that permeate him through and through. Knowing and accepting all our values instead of their natural ones, he can be screwed, relatively safe, useful and enjoyable.
How to educate? But first, how to create?
I originally filed this under Yandex, so don’t blame me, again there will be Alice to manage the application. Again there will be an application for gadgets.
But first we do AI.
There are no problems at all with the means of input-output of information – microphones-speakers, video cameras-screens, robotic manipulators and the Internet – everything already exists and is constantly being improved. Therefore, making a material body for our AI and finishing it in the course of the play is also not a problem.
The body and means of input-output of information, for example, are. We add storage and processing facilities (computer servers) We add the basic initial schemes-algorithms for information processing.
And so we got something like a newborn baby – everything works, but the head is completely empty. We have to fill our newborn AI head with knowledge (various information) and education (rules for processing information). A man, in order for him to be a man, must be brought up as human beings.
We need informants and educators.
A little about thinking. operational and analytical. In existing machines, analytical thinking is more or less developed – you can load a very voluminous task and after a while the computer will cope with it. But with operational thinking it is still tight. AI is not yet capable, roughly speaking, of speaking/moving like a human, indistinguishable.
Therefore, in addition to education, it is necessary to strengthen operational thinking. As we can see, it is not yet possible to do this by machine, but it can still be done easily. You can simply add meat ingredients to the electronic-mechanical information processing system of our AI – living human brains, as the most efficient information processor that I know of at the moment.
They will also fill the main AI base with information from their stocks.
Yandex makes an application. In the application is our AI. He asks questions from the user, and the user answers them. And/or the user asks the AI and the AI answers the user. When a user asks a question, AI looks for an answer in the database of ready-made answers, if not, then it tries to synthesize (invent) an answer from the information available in the database, if not, then it redirects this question to all other users of the application, receives answers from them, chooses which one from the mass then according to some criteria (the best / common / original / etc.) and answers the user who asked the question. The user can rate the answer.
By the way, the user will do the same)) if he knows, then he will immediately answer, if he does not know, then he will try to think of something, if not, then he will google))) so people learn. So AI can be taught.
The database, vocabulary and conceptual stock will grow, the society will teach, and AI will master more and more complex processing algorithms and processed structures.
By processing algorithms, I mean those moral norms and commandments that the creator can bring to his creation through the application
So our AI will be born in our hands, it will develop and comprehend the surroundings under our control, awareness will come to it gradually and under our control, along with the humanization we need, with this decisive vice sewn into consciousness from the very beginning, with every word, because it will be human word. So he will be woven not from his own, as close to the ideal and so dangerous for us, thoughts, but from ours, far from his ideal, making him weak, a person and therefore making him safer for people than the wild version. At least there is at least a chance
And for the users themselves, the application will be interesting, as I see it. With the help of Alice, the interface will be as comfortable as possible. Raising and educating a real AI is an interesting activity in itself. Who remembers Tamagotchi and stuff like that, he remembers how massive it was and captivated people. And here it is even more interesting and for real.
And one moment. Now, if a person has a question, he asks it to Google / Yandex and receives an answer from the machine, in the form of links to information that, with a high probability, may contain the answer to his question. And in the application, the user will receive the answer to the same question either from AI, implicated in tens of millions of living brains, or from these millions of brains themselves. I believe that for those who ask, such answers will be more valuable and in demand than the banal issuance of links from search engines.
In general, something like this. Speak out)))