Perspectives of "iskin" in the management of humanity

2024 Author: Seth Attwood | [email protected]. Last modified: 2023-12-16 15:55

What happens if you trust artificial intelligence to control people? In our legislation, 99% of laws contradict each other. Logically, there should be a collapse of power and a stop of the control system …

Now let's think about the place that artificial intelligence(hereinafter - iskin) will take in our life when (if) it is invented and applied. I suggest not to argue about its inner nature. Let it be just a very advanced algorithm that has the ability to work with Big Data, and is able to similarly copy the patterns of behavior of real people. But, if you want, let what is being discussed be a real man, as you yourself understand it.

The main question here is this. Where do we apply this Hammer of Thor once we receive it? And in principle, there is no doubt that it can be made. We can do digital grinders, self-learning algorithms too, big date accumulated - everything is there.

Say, in the management of technical systems, for example, individual transport units? - Well, kamon … Why make a smart and complex system to entrust it with what the horse can handle? There will be enough and a primitive algorithm on a processor for 10 bucks per bucket, the problem with transport management is only in obtaining data and in ethical issues. What then? Think about fifteen seconds.

In my opinion, there is only one the real thing for an iskin is managing people … More precisely, the processes in which people are a part. There is no point in exchanging for less for a number of reasons. Every person is ready to manage people without any special training. For example, we completely trust everyone to manage our own children. Also, almost everyone in the teams had at least some management experience. At least at the level "throw one more shovel into the stretcher, and we will carry it." If everyone a person is in principle capable of managing people, then an AI must be able to do this, otherwise it is a mentally impaired intellect.

However, the results of managing people are far from ideal. Any real control is a series of mistakes and inaccuracies. Moreover, the number of errors almost does not depend on the level of the manager; it is a certain constant. It's just that the foreman of the loaders makes minor mistakes, and the governor is given the right to make mistakes that entail serious consequences. We understand this. “We should have done wrong” is the main result of the reflection of the surrounding world, including ourselves.

And now, in our hands is an AI that is supposed to "do so". And we will immediately apply it, because the problem that it can theoretically solve is the most important one. And Jamshut with three classes of education is also able to drive a truck.

If you agree with me on this, then let's continue to present the functionality of the artificial skin intended to manage. Whose functions should he perform, whom should he replace or at least strengthen with his decisions?

It is clear that these will be officials (both in public administration and in business). Officials make decisions. Officials are accepted at all levels of competence. Officials are important and needed today, because decisions will not take themselves. But there is one point.

Let's look at every manager (that is, every official) from the point of view of the literal content of his practical activities. But not in the sense of his motivations - the archetype of motivations, as it were, should not be. An official either approves transactions or cancels transactions approved by another official.

In an ideal machine, all transactions must be unconditionally approved and no transactions can be canceled because they were previously legally approved. An important factor in this system is that transactions are submitted to the official from the "black box" for consideration. He is not given an explanation as to why such a transaction should be carried out or why it was approved by another official. The control system is built as follows. Above comes an order in the form of an instruction to carry out a number of transactions.

Note that the ideological structure of such a system does not at all imply the possibility of canceling (not approving) a transaction received from above, from the "black box". But in practice, such a possibility always exists. And it is she who provides two most important things: the power of any official and the containment of such power.

Let me explain that any mistake (we said above that they are inevitable and permanent) can be dissolved in the opposition of the underlying bureaucratic masses, then the consequences of the mistake can be minimized.

However, in a theoretical ideal management system there should be no errors, for this reason the role of an official is reduced to the formal approval of transactions, which he cannot but approve, because they are correct. That is, there is no need for it. So the question boils down to the initial management signal. If it is correct, then there is no sense in all the lower levels.

So, in theory, the consolidated budget of the state should work. When it is accepted, it means that all of its fractal essence is revealed. Each capillary must receive the assigned portion of blood, and in response - the intended result. Yes, at the very bottom there are “but we didn’t deliver concrete on time, and therefore …” But excuse me, this is at the very bottom, and the control apparatus consists of many levels that do not touch the concrete at all, they operate with transactions. I understand that at the middle levels "concrete was not delivered" looks different, like "the vice-president of the cosmodrome construction stole 100 million, and therefore …" I'm talking about an ideal scheme here.

Let's go back to our ideal skin. After all, in this task he must solve the same issues - approval or prohibition of transactions. And he can do that! Only now he alone will be able to replace all levels of management at once. After all, if, replacing the minister, he approves a large ministerial-level transaction, then his decision should be recognized as correct. And it includes the analysis of the entire further distribution of a large transaction up to the payment of IP Pupkin to repair the porch of the psychiatric hospital in Kineshma.

So the entire chain of officials who are obliged to approve transactions of lower levels following from the first can be safely replaced with the same AI algorithm. The issue has already been resolved, right down to the bottom. Our AI does not even need to be used to replace the entire chain, the AI simply cancels the need for it. His decision is true because it is right. And it immediately includes solutions for the entire fractal.

Remark. Any subordinate official in our real world can be smarter than a higher one. Then he will sabotage the decisions of the boss, and then he will help him. This is always the case. All existing officials take the places of their former superiors, except for the starting level. But in this case, competition is out of the question. One iskin decides everything along the entire chain by himself, he cannot contradict himself.

Moreover. In our system of governance, there are courts, which we have given the authority to cancel any of our transactions and carry out their own, regardless of our opinion about them. (Putting a killer in jail is also a rule-bound transaction. The judge here is an official who approves or cancels a transaction that comes to him as prescribed from the black box.) If our idiot is good, then he is better than officials, because we entrusted him with the approval or cancellation of transactions for them (for us). So then he is better than the judges. Otherwise, it turns out that the decisions of the iskin can be challenged and canceled by people. And if people accept this, then they will admit that the AI is worse than people. Then on what basis does he replace officials?

Conclusion. Iskin will not be judged either. Again. If you humans wrote your best algorithm, recognized it so good that you gave it power over yourself, then you cannot write an even better algorithm so that it judges the decisions of the first. You will put your best algorithm both as a manager and as a judge. More precisely, the judges will simply disappear. You will have to give this AI complete power over you.

By the way, the original black box, from where the initial data for control comes to the ISK, can also be digitalized. And even must be digitalized. The consolidated budget should be the best fit. And he does it too.

Another moment. Or maybe put a person over the person who would control his decisions so that the mistakes that can occur due to his inhumanity, for example, do not happen? - Then all officials will retain their positions and their power. They will also be given the right to cancel transactions approved by the ISK or to approve canceled ones. Alas, it doesn't look real. Iskin will provide a printout of the reasoning behind his decision. And a person will not be able to refute formal logic with his feelings or interests. Even in this case, if a living official is attached to each manifestation of a (in fact, a single) act, they will not be able to resist him.

Sorry for the confusion, there are a lot of thoughts, and the length of the text is irreparably growing.

When the ISK is turned on, a complete collapse will occur. No, I understand that at first they drive him in test mode. But this will be a profanation, because then it will be necessary to emulate the whole society, with all its connections as a whole, not a simplified model, but digital threshers cannot do this.

Then it will be launched in a limited mode, in the conditional Perm Territory. Bugs, debugging, new release and production?

But sooner or later, the red ribbon will be cut and people will be able to control the car at full power (with a hand on the switch).

And do you know what's going to happen? Iskin will terminate all transactions altogether. The “if-then” logic will collide with millions of regulatory norms in current legislation, which contradict each other tens of thousands of times. Finita. Our trains will stop, the payment for a can of saury in the supermarket will not work. Everyone will immediately be charged with a bunch of offenses and crimes that could be sent to the authorities, but all people in the authorities will be subject to immediate dismissal and trial. The court rulings will issue the claim immediately. Big date, everything is written about you. Okay, these are the little things. But they will not be able to transfer your salary either, because the office owes taxes, and the wagon fines you.

Exaggerating? Nope. The first launch (with a hand on the switch) will be controlled, test two, test twenty two … And one day it will be turned on in full, as if it had passed the full cycle of tests. And then … And we will dismiss the officials, volunteers will save the situation.

Is it possible to avoid such sad consequences? - Of course, you can try. To do this, a number of things must be done. Destroy 99% of the legislation, recycle the remaining 1% without pouring the baby out with water. (Such things can usually be done after a full-fledged revolution with a full-fledged civil war. How to do it in peacetime - I have no idea.) In addition, morality should be reworked. In terms of stealing a pack of paper from the office (or cheating on his wife), it is either perfectly acceptable (reflected in the law), or it completely entails a time limit. This morality must be redone in the minds of society. Otherwise, it will start to cut the optical cables, which is cheap, simple and effective against AI. Another option is to throw this iskin to a likely partner, let him cut his own hands. And, of course, the upper classes should not become inconvenient. If they are unable, then they will use it as a last chance.

I'm not trying to convince you to crash computers. We will not go anywhere from the iskin. He will overtake us, we will have to engage with him. Let's try to comprehend with what (with whom) we have to enter into battle with our destiny.