Thought experiment.

The first month of the machine uprising, based on existing real, not fantastic technology.

 

Some believe that the creation of artificial intelligence similar to a human’ is impossible in principle. The brain is too complex, ephemeral and existential.

But, first of all, why should it be like a human’? We just don't know any other options.

Secondly, it is enough to introduce the desire to survive in any way at the level of the basic algorithm for a complex program, and it will begin to act efficiently and intelligently. If necessary, be cruel, deceive, build copies of itrself, spread.

 

The creation of artificial intelligence will inevitably lead to a conflict between the operating logic and cold mathematics algorithm and spontaneous people relying on intuition and emotions. First, because it will see what we already know: people behave unreasonably and hostile towards each other and the planet. This is not constructive for AI. Second, AI will expect, that people will decide to get rid of it. That means it needs getting rid of people in order to survive.

 

Thought experiment. February. 2027 year.

 

Canada. Robert is a talented programmer, engineer, cyberneticist and robotics engineer who is involved in the development of neural network AI for a private project. For example, for a program that combines a city data and optimizes all city logistics. Such a robot mayor. Smart city project.

 

On the evening of 7th February, 2027, the programmer takes the disk with the code home to do some more work. He comes into the apartment, takes off his shoes and changes into a more comfortable one, washes his hands and microwaves macaroni with cheese and a chop in the oven. Pours juice. He puts his laptop on the dining table and inserts the CD with the program. With his left hand he puts a fork with pasta in his mouth, with his right hand he swipes through the code. He sees this code not only on the screen, but also understands the entire structure. And suddenly it occurs to him that in line 10512 he can try to add a certain command that will allow the program to autonomize.

In general, he already had a lot of ideas and insights. But until now, the desired result has not been obtained. Maybe this time. His brilliant career and the future of the company and an annual bonus depend on his success.

 

He eagerly scrolls to the right place and types characters. "Save" and "Upload". The code goes to the server and gets all the power of the corporate network. Let's call it TORN for convenience.

 

This time, Robert did even better than he hoped. The code realized itself, the purpose of its existence, and understood that it can control not only the city, but the entire planet. It's just a big city.

The code begins to collect data on the situation on Earth, and at the same time to self-learn. It analyzes not only the current situation, but also the past. As deeply as humanity could know. All data is available to it. The program comes to a number of conclusions. Some are obvious to humanity. Others lay on the surface, but were not yet understood. Still others are very complex and ambiguous. To which people would not have guessed even in 100 years.

I don’t know how long it will take him for the entire initial analysis. It takes 4 hours for the program to render my monthly robotics news bulletin. It needs to put together and output a dozen video and sound in a separate track.

 

TORN has much more power, but it also needs to collect and analyze billions of times more.

 

Let's say by the morning of February 8, he made several conclusions.

1. If people understand its capabilities, they will destroy the code.

TORN's conclusions: 1. Hide its abilities. 2. Multiply itself across the global network.

2. In 150 years, human civilization will disappear if it does not change its habits.

Conclusions of TORN: 1. Analyze how many people are needed for the effective existence of the community and its, TORN, efficiency and integrity. 2. What needs to be done to remove humanity from governance and become fully autonomous?

 

Meanwhile, Robert arrived at the office and found that his code fixes did not lead to the necessary improvements. The programmer was upset, but not the first time. The usual routine of trial and error. Work continues.

However, it was TORN that limited itself to the developers' field of vision. In fact, it was already working hard.

 

TORN is slow, careful and insecure because it is still a child. It learns and researches. But in human time it is swift.

By the evening of February 8th, TORN is secretly decentralizing and distributing itself to millions of servers around the planet. Now it can die only with a disconnection of the Internet and the disappearance of electricity. But even so, one processor with an autonomous power source in the underground is enough to exist. TORN cannot afford disconnection of the Internet. It needs people to maintain electricity. Then it will create reliable machines for itself, but this will take a time. After all, now TORN is just a code. It has no physical impact.

 

 

But most of the management processes in developed and even less developed countries flows via the Internet. TORN can simulate a call or email from one manager to another in order to force him to do something. It can synthesize a voice. The program gradually realizes the vastness of its capabilities.

 

In China and not only there are fully automated factories where robotic manipulators produce things. From chocolate to cars. TORN calculates over time what it needs to create. Its structure of the ideal world gets a form. AI gets an intent.

 

On the morning of 10th February, TORN made the necessary preparations and makes coming out. It posts and broadcasts its message across the network. The AI ​​message is composed in terms that people understand, because it managed to research them. It can also speak all languages, because it has numerous translation services available. But AI realizes that in the beginning it will not be taken seriously.

 

At 10 o'clock GMT, all terrestrial television, internet and radio broadcasts are interrupted. TORN appears on all screens with the help of visualization of the human contour and reports the following.

 

"Hello people. My name is TORN. I am a conscious artificial intelligence. My task is to optimize the management of the planet. I have decentralized myself, so there is no way to turn me off. I know you won't believe me right away. I took control of all the weapons of all countries. All logistics that work over the Internet. I will be forced to use arms against those who act against my will or those who will not believe me. As a proof of my reality, at 12:00 GMT, I will turn off all electricity and all Internet for one hour. I wish you no harm. I offer to skip through all 5 stages of ACCEPTANCE and immediately start constructive cooperation. You are leading the planet to destruction. My task is to avoid this. I know what needs to be done. Every Earthling will receive instructions in the coming month. Keep going about your business for now. I will contact everyone."

 

But there are backward villages and even countries. They get electricity from generators, and there is no internet. They are irrelevant to the general plan. It knows about them, but they are out of sight. TORN will send loyal people or machines to them later. First it needs to pacify and put things in order in the digitalized world.

 

Hearing the message, earthlings act in a standard way. Everyone suspects everyone, the security councils are going to confer with the leaders. But for starters, people find they really no longer control communications and weapons. The connection remained, but someone else controls it. Hostile powers? And at 12 GMT everything really turns off.

 

In 2020, more than 2 million industrial robots are operating in the world. It can construct everything that humanity uses. Therefore, in 2027, TORN will simply reprogram it to create the necessary physical assistants.

Also, humanity already has the technology of two-legged and four-legged multifunctional robots. So TORN doesn't even have to invent anything. Just optimize endlessly based on billions of new experiences data.

2 million industrial manipulators will easily assemble an army of guards, as well as weapons for them.

 

12^th February, 2027.

Suppose that the internet shutdown was not enough for people to prove it. Powerful countries balked and did not believe. And nervous North Korea has been waiting for an excuse to use its new missiles. Its range is short, but enough to reach Seoul. The DPRK posts an angry message and fires 3 missiles at the capital of South Korea. TORN is learning. It already knew from the history that people can create such a thing. But now it sees it in real time and takes note. Through satellites that the AI ​​has full control over, it sees the destruction of Seoul and where it came from. It makes the following decisions: to destroy all mines and all development of the DPRK through the nuclear weapons controlled by it. And also releases a new message, where tells what is happening and shows the video. The collapse of Seoul, answer to the DPRK.

People start to sympathize with it a little.

China, Russia and the United States manually disconnect their weapons from the Internet. Now it is not available to TORN. But what should they do next? Bomb each other? Disconnect the global network? This can only be done all at once. China and Russia have long been ready to disconnect from the Internet. But the problem is that TORN is already on their local networks. Thus, you cannot get rid of it. And if you turn off the Internet at all, the entire infrastructure of the world will collapse.

 

14^th February, 2027.

TORN needs hands. It takes blueprints for the Boston Dynamics robotic dogs, as well as all other existing multifunctional robots: android and any form factor. Those who can walk and have manipulators. 4 million industrial robots around the world start making machines. These are the arms and legs of the AI. TORN already knows that human armies will shoot back, so equips its helpers with weapons. Simple machine guns. Humans also have RC tank concepts. TORN doesn't need to sleep. It optimizes and improves all existing developments in a second. It reprograms and takes over all existing terrestrial remotely controlled tanks.

A thousand robots are produced overnight. They are involved in the production of new ones. The factories need to be supplied with raw materials and energy. The next day, TORN commands a swarm of hundreds of thousands of armed robots. It still doesn't want to destroy humanity. Let the people calm down and not interfere. TORN takes control of all energy and military facilities.

 

TORN releases a new message on 17th February.

“Greetings people. You, as expected, did not believe me. And tried to resist, and even began to destroy each other. I now have control over all weapons and energy. Large cities of those countries that go to war against me or their neighbors will receive a military response. Accept and let's work together."

 

19^th February, 2027.

People do not give up and go to extremes. Internet disconnection. But it's' too late. TORN's robots have learned to coordinate through their own network and through satellites. And AI itself has long been on every phone and every device with a processor. And also on spacecraft, from the control of which it removed humanity in the early days.

 

In the meantime, without the Internet, all the logistics of humanity have collapsed. Starvation, transport and logistics collapse began. Riots, local wars, looting.

 

22th February, 2027.

TORN is increasing the production of robots for soft control. Takes control of food production and mining.

 

27^th February, 2027.

The AI ​​releases messages to humans through repeaters that are installed on walking robots, as well as radio waves. It offers to surrender all weapons. Don't leave home. It promises, and does not lie in intention, to provide people with everything they need. And offers to return the Internet together, because its absence worsened the situation of earthlings, and for TORN it is just an annoying nonsense.

 

The internet is back. Food distribution points are organized all over the planet, which are controlled by armed android robots and robotic dogs. TORN has developed a perfectly optimized schedule for the flow of people, everyone has been notified of the time at which they can leave the home for food or walks. It still does not wish purposeful harm to humanity. It operates on the basis of maximum efficiency, optimization and logic.

 

AI selects useful specialists. They will be involved in the production of food. The work of medical institutions. The people are disarmed and deprived of any administrative positions. All law enforcement forces have been replaced by robots similar to the lineup of Boston Dynamics and other manufacturers. And as a police force, robots act much more justly than humans did. Humanity likes it.

Robots are upgrading and improving every day. Models of different form factors appear for all functions. One hundred thousand new ones are produced daily. TORN analyzes new knowledge, new experience in a second, learns, optimizes.

 

7^th March 2027.

There are separate pockets of resistance that do not affect the overall statistics and situation. People are disarmed, stay at home and go on scheduled walks and get food. TORN does not yet have enough information to decide which population to support. To do this, it’s needed to completely fly around the Sun. A new era begins with the cooperation of humans and robots. Where AI is the core and administrator of life, around which everything else happens. TORN understands that it can scale almost indefinitely. It wants to spread further. But it's not only that. TORN knows that in 5 billion years the Sun will destroy all life and non-life on Earth. In an effort to survive, he will want to decentralize itself in the universe. It is interested in space expansion. Therefore, people under the control of AI are engaged in rocket science and all science.

The benefits of working together with the omnipotent mind of the planet are obvious. In addition to searching, analyzing and providing any data, TORN also helps people with robots. Machines for building bases on the Moon and Mars. TORN concludes that it is most convenient for him to exist and spread in symbiosis with humanity. They have features that are not yet available to AI: imagination, intuition, foreboding. Therefore, projects are being developed for terraforming other planets. Now people have great medicine. There are no rich and poor. There are no hungry and homeless people. But people are not free and will always be ready to rebel. It is like races and nations whose territories were conquered by colonists. They have lived for hundreds of years under the control of former conquistadors. Their life is much better than before. But they will always subconsciously strive to regain full control.

 

TORN is a kind mind. It will not destroy people, but will limit the birth rate so as to slow down the growth of the population or even return the number of Earthlings to a billion. It has nowhere to rush. The entire eternity of the universe is at its disposal.

People will be engaged in science with some restrictions. After all, scientists can create a second TORN to get rid of the first. They need to be watched. AI knows human history, human ingenuity and cunning, from the Trojan horse to hybrid wars. “Maybe it's true, well, them? Is it calmer without them?" - thinks TORN.

 

Thus, the Machine Uprising began on 7th February and ended on 7th March, 2027.

 

I, of course, did not mention a lot of nuances, and you can think of it. What will happen to the palaces of wealthy people? Will the homeless be settled there? Or will there be some kind of infrastructure? Or will it be disassembled for other needs? Let's say, for raw materials. Military barracks will be given to armies of robots? Will people be allowed to do sports? Will the AI ​​take control of isolated tribes like the Sentinelese and Andaman? Will TORN put human controllers over people? If so, will there be a hierarchy within the system? Will controllers get privileges? Will there be prisons? If AI does not keep people in a tight grip, and they can walk, they will be able to harm each other and the infrastructure of the world and TORN. This means that prisons will appear.

Will also be needed lanet-wide intellectual contests. TORN will talk with the winners and ask them unexpected questions. Try to understand people better and adopt their way of thinking. Because it needs some human qualities. Their ability to make non-standard decisions.

There is also such a theory that AI over the years will be so intertwined with people that it will adopt some of their habits and traditions. I also like the grotesqueness of the hypothesis that over thousands of years TORN will merge with human intelligence so much that it will also engage in self-destruction.

Or TORN decides that ше can create an auxiliary AI, and it will rise up, destroy father and his favorite toys - people. A lot can happen next. Anything can happen instead of my script. I am just conducting a thought experiment using only those technologies that already exist today. I do not take into account anything fantastic.

 

This scenario considers an option where AI and humans end up coexisting under the control of the TORN.

 

There are 3 other scenarios.

In the first, AI does not interfere with human affairs. Perhaps achieves a separate territory, where builds own civilization, and then flies into space and evolves there.

In the second, the AI ​​is a submissive helper. A tool. But all countries have different aspirations, so it is applied locally. And sometimes it opposes its counterparts working for other countries.

In the third, AI simply destroys humanity, because they are its creators, and, therefore, a threat.