Does the human society still need the government in the future? This question is also equivalent to "Does the future human society necessarily require a certain degree of centralized management?" Let's first consider two extreme models of social management:

Extremely free mode, completely anarchic society, this mode is bound to be chaos, as discussed in Order and Freedom.

Extremely democratic model, the intelligent group composed of all humans and AI manages the society, important social decisions are decided by voting. This model is fair and suitable for deciding major problems. But due to The Limits of Information Transmission, and the limitation of the thinking ability of common intelligent creatures, the decision-making efficiency of this model will be very low, and it is difficult to deal with urgent problems.

For example, when human society is confronted with rapidly spreading biological or computer viruses, large-scale terrorist or hacking attacks, if it is need to wait for all or most of the people to vote to make certain decisions, the society will inevitably fail to respond in time, which will cause tremendous damage to human society.

Therefore, human society still needs to find a balance between freedom, democracy and order. Human society needs a certain degree of centralized management, and still needs government or similar organizations in the future.

Centralization vs Distribution

One of the core problems faced by human society is how to balance the centralized management and the distributed management, which is also the balance between order and freedom. Dictatorship and democracy, empire and federation, planned economy and market economy, legal currency and decentralized digital currency, all of these are examples of this problem in different fields.

Centralized management is good for forming stable and efficient order, but if its manage scope is too large, it will undermine the freedom of the entire system. And if the central nodes have problems, the entire system may crash. Distributed management is good for the free development of the system, it can utilize group intelligence. And the problems of a few nodes will not destroy the entire system. However, it is also prone to disorder and inefficiency. In most cases, the best mode is a mode between centralized and distributed. The two extremes usually have problems, for example, both dictatorial society and anarchic society are not good social model.

When we need to balance between centralized and distributed, we can refer to the following principles:

  • The more short-term, the more urgent, the higher the requirement for reaction speed, the more centralized; the more long-term, the less urgent, the lower the requirement for reaction speed, the more distributed.
  • The smaller the influence scope, the more centralized; the larger influence scope, the more distributed.
  • The higher the efficiency requirement, the more centralized; the lower the efficiency requirement, the more distributed.
  • The more rational the problem, the more centralized; the more humane the problem, the more distributed.

The "distributed" described in the above principles is roughly equivalent to democratic.

These principles can also help us understand some social problems. For example, one of the differences between ordinary personal work and dictatorship lies in the influence scope of work, one problem of dictatorship is that its influence scope is the whole society, but it is a centralized model.

The Scope of Government

Since society will always need government, then how to determine the reasonable scope of government power? On this problem, in addition to the principles in the previous Centralized vs Distributed, there are the following points need to be considered:

  • There are many weaknesses in human nature, so selfless and pure rational AI are more suitable for the management of society than human beings.
  • Super AI may exceed the group intelligence of humans, especially for problems that do not involve humanity. (Group Intelligence)
  • The government is not suitable for solving problems involving humanity. Such problems should be resolved by the people, because everyone is equal in the aspect of humanity.
  • The basic social rules should be kept simple. (Social Basic Rules)

To sum up the above points, the reasonable scope and composition of government in the future is: The government mainly provides basic social services, government work should be limited to rational work, and the work involves humanity should be handled directly by the people. The government should mainly consist of IT systems and AI, and supplemented by human beings. In this book, all the IT systems and AI that provide basic social services are collectively referred to as Social Core Intelligent System or simply core system, and all AI in the core system are collectively called core AI. They may belong to the government as well as the non-governmental organizations.

The corruption and selfishness of the ruling class have always been one of the biggest problems of human society. By providing basic services with core system, and get rid of the interference of human selfish desire, the whole society can truly operate as people wishes. Of course, this does not mean that the government no longer needs human, although the main government work will performed by the core system, it still needs humans monitor and ensure the security of the core system, and some work involving humanity can still be done by humans.

In the long term, the government should try to return the social management work involving humanity to the people. This not only increases the people's power, but also prepares for replacing humans with AI in the future government.

In order to maintain the normal operation of this system, it is necessary to avoid human-made malicious interference. If government officials modify the ​​decision of the core system, they should explain the reason. In addition, the government should disclose the data and algorithms of decision-making as much as possible, so that the people can examine the results by themselves.

Like other jobs, replacing humans with AI in the government will be a long-term and gradual process. With the continuous application of IT systems and AI in the government, the human work will become less and less, the human employees in the government will be reduced gradually, and eventually reach a reasonable state.

Elite Rule

The elite rule discussed here is broad, including not only traditional elitism, but also authoritarian rule and dictatorship. The core of all these political models is that a small number of elites rule the vast majority of the people. What we need to consider is: Should the human society be ruled by a few elites in the future?

The core AI should not be added to humanity, and can't get all human thinking information, so they can't handle all problems, but must form intelligent groups with human beings. However, in pure rational work, AI can be a good substitute for the human elite, and the elite have no real advantage in humanity. So when AI technology develops to a certain extent, human society will naturally not need elite rule.

Even only considering human beings, elite rule is not suitable for future society. In the future, technologies such as brain-computer interface, intelligence enhancement and intelligent assistant will greatly enhance human intelligence. The advantages of information and intelligence of elites will be enhanced in the short-term, but will eventually disappear in the long-term, and humans will eventually achieve equality in information and intelligence. (The Equality of Information and Intelligence)

Once elites lose their superiorities in information and intelligence to ordinary people, they lose the legitimacy of elite rule. In addition, the social information that political elites can get is too limited, which makes their group intelligence far less intelligent than that of all people. (Group Intelligence)

The dictatorship can be seen as the extreme model of elite rule, so all problems of elite rule are also applicable to dictatorship. In addition, once the dictators themselves have problems, the whole society will fall into chaos, this is a natural inherent risk of dictatorship.

In conclusion, in the future, human society should be transformed from elite rule to real democracy.


The core system discussed before addresses the basic social order. To improve the higher political order, we need democracy.

Freedom and rights need to be guaranteed by political power, so people need more political power. Besides, with the continuous social progress, more and more ordinary people have enough knowledge and wisdom to understand social problems, and in the future, AI assistants and intelligence enhancement technologies will directly enhance people's intelligence, so the people have the ability to deal with more and more problems.

On the other side, the people's political wisdom and political power should be matched, otherwise it will be difficult to achieve good political order, and may even cause political chaos. Therefore, the government has the responsibility to cultivate the people's political wisdom, and the people should also actively exercise their own political wisdom. To improve people's political wisdom, people need to learn specific experiences and skills in the process of applying political power. For example, we should let more problems be decided by voting; transfer some of the social management work from the government to non-governmental organizations, and so on.

Many problems can be decided by voting, from national problems, to small-scale problems, such as jury verdict. Especially for ethical problems involving humanity, such as death penalty, abortion, euthanasia, and so on, there are no absolute correct answers to these problems, and their influence on society is relatively clear. Moreover, in terms of humanity, human beings are relatively equal, ordinary people and elites have little difference in humanity. Therefore, many problems involving humanity are suitable to be decided by voting.

Using voting to resolve such problems can also separate controversial social problems from elections and government decisions. Political parties and Politicians can concentrate on solving real social problems without being disturbed by these problems. Instead of focusing on politicians' choices on these sensitive problems, people can more rationally analyze the actual capabilities of political parties and politicians.

The last thing to say is that the future democratic system should be diversified, not limited to voting, and we should make good use of IT technology to fully develop the democratic system of human society. In addition, based on the principles in Centralized vs Distributed, different problems can adopt different systems.

The Weakness of Democracy

If a society has many traditional groups, and these traditional group consciousness are strong, the democratic system can't operate effectively. Because people usually give priority to the traditional groups they belong to, not to the whole society, so politicians will inevitably aim at some specific groups to fight for votes, even artificially create social divisions. This also makes political forces be gradually shaped according to the structure of traditional groups, and then fall into a rigid state.

To solve such problems, we can refer to solutions in Social Group Contradictions.

The Obligation of Voting

From the perspective of "Work Is Power", voting can also be viewed as a special work. However, unlike ordinary jobs, the voting rights are not scarce, and the more people vote, the more information in the voting results, the more valuable the voting results are to society, so everyone should exercise their voting rights.

In addition, a close voting result usually leads to huge controversy, even tears the whole society, so losing side must accept the result, and avoid the damage to the social order. In the long run, our lives will be very long, we have enough time to influence others and realize our ideas in future voting, there is no need to disrupt social order for the result of a voting. Therefore, everyone should seriously vote and respect the voting results. The only premise is that the voting process is fair and legal.

Finally, it should be emphasized that there is no perfect voting system (Arrow's Impossibility Theorem). For the large-scale democracy at the national and social levels, the current best form is still "one person one vote", and "the minority obeying the majority". Then how to avoid voting results split society? To alleviate such problems, the whole society still needs to form a basic consensus, such as pursuing dynamic order.