A typical and current concern about artificial intelligence is that sooner or later it will somehow overtake human society, dominate the wo...
A typical and current concern about artificial intelligence is that sooner or later it will somehow overtake human society, dominate the world, and ultimately, as an existing unnecessary problem, even destroy the whole of humanity. Although this concern is real, because it is potentially possible to create an artificial intelligence that can behave in a similar way to us humans, who not only destroy less capable living beings according to our desires or interests, but typically we also destroy each other, this type of actual concern about artificial intelligence, however, is probably unrealistic. It is unrealistic not only because the artificial intelligence that we can create today, according to our present knowledge, does not have an independent will that would make this dangerous motivation possible, but also because the artificial intelligence that we create did not arise under natural evolutionary conditions, so that the life instinct that ensures self-preservation does not necessarily develop in it, which would make the destruction of other things to sustain itself rational for it, and intelligent thinking without life-sustaining natural instincts may much prefer symbiotic coexistence rather than the destruction of created existing values. At least we can hope for this.
Humanity probably faces a much greater and more immediate danger in relation to artificial intelligence. The greatest danger to humanity's existence in relation to artificial intelligence is that humanity has come into possession of a device that is more intelligent than any human being, and potentially even the whole of humanity, that has no will of its own, and that has no intrinsic, necessarily present moral constraints.
Apparently, today's most advanced artificial intelligence already possesses more information than any human being, since it is potentially capable of absorbing and retaining any amount of information, and we, humans, provide all we know to artificial intelligence with the opportunity to learn from it to become a useful tool for us. Furthermore, today's most advanced artificial intelligence potentially can have more advanced knowledge than any human, because it can discover correlations among data from the information it possesses at a level of sophistication that far exceeds human capabilities, and it has no potential limit to the computational power that makes this function possible. As a result of the above, artificial intelligence is potentially capable of determining hierarchical relationships between the correlations it discovers in a way that exceeds human capabilities, and potentially able to applying them to achieve given goals, i.e. artificial intelligence is potentially capable of reasoning intelligently to an extent that exceeds human capabilities, and potentially for all of humanity.
However, it can also be observed that in the use of the currently existing artificial intelligence, at the level of its development, at its ability of intelligent thinking, i.e. in recognizing connected correlation sequences from the data at its disposal, at any level of knowledge that the currently existing artificial intelligence can have, in the use of these intelligent abilities, even seemingly independently for any kind of task, somewhere in this process, typically at the beginning, there is always human beings present for defining at least the primary purpose. At present, there is no artificial intelligence application where a human being does not play a decisive role in the process of intelligent operation, in determining the goals of the intended process. It can be stated that the current artificial intelligence is not capable of directing intelligent thinking completely independently, exclusively by its own intention, which exists by itself.
While this characteristic may sound reassuring in the context of currently deployable artificial intelligence, it also implies that artificial intelligence, in the absence of its own self intention, does not have its own independent value system either, i.e., a self-derived moral regulator that could somehow limit its operational goals. In the case of humans, this limitation was primarily formed by social-evolutionary development, but artificial intelligence does not currently have such a feature. There are no evolutionary constraints on the development of social laws governing proper human coexistence for robotics, which would naturally regulate the behavior of artificial intelligence. The behavioral constraints of artificial intelligence are merely the rules that humans define for it. However, these constraints depend entirely on the human intention to create and use artificial intelligence.
The need to regulate artificial intelligence has long been proposed in science fiction literature, but this regulation would actually mean restrictions of actions for humans, who have evolved under evolutionary conditions, typically possessing a desire to control for expressing dominance, and who also want to use artificial intelligence as a tool. Without social coercion, humans are reluctant to limit their possibilities in the use of artificial intelligence. The desire to rule is also a well developed, typical characteristic of the leaders of society, and even more so of the leaders who tend to authoritarianism, especially for those, who also have the most decisive and exclusive role in forming the social order, and who want to and are able to enforce their desire to rule with the greatest potential.
It is not surprising that authoritarian social systems not only promote the use of artificial intelligence without restrictions for them, but also widely use such tools. Artificial intelligence, by its very nature, is an ideal tool for controlling society, which is a necessary condition for the existence of authoritarian regimes. Artificial intelligence is capable of continuously observing the behavior and communication of members of society, discovering correlations in it, and, based on an external, determining human intention and without self-existing moral constraints, even independently and actively performing actions in accordance with the goals defined by its controlling humans.
With access to the behavioral and communication information of people in society, artificial intelligence will be able to filter out not only those individuals whose behavior is apparently dangerous to society, or even more importantly, in the case of authoritarianism, dangerous to the ruling power of governance, but through access to communication information channels, according to the guidelines of the controlling people, in the absence of social restrictions, the governance will also be able to distribute information automatically generated by artificial intelligence, according to the guidelines of the governance, which can manipulate the behavior of the members of society and the will of the people, even in a personalized way, by influencing them as the needs of those in power (those who can access all information and those who can control artificial intelligence) require. Artificial intelligence is an ideal tool in the hands of governments that manage human communities, even by spreading false information, in the best case for maintaining social cohesion, but also in the worst case to extend the power of the rulers in space and time without limits, and thus to maintain the social order for those in power in a way favorable to them.
If we consider the two possible stable equilibrium states of social orders, one being the democratic system, where the will of the majority of the community with the recognition of general moral norms prevails, and the other being the authoritarian system, where the will of a narrow elite prevails in the management of society, even without considering generally valid moral norms, then it can be stated that from the point of view of social order, the democratic system is an unstable equilibrium state, and the authoritarian system is a stable equilibrium state of social order. In the case of a democratic social order, a small deviation toward autocracy, which may simply occur due to the nature of the functioning of democracy, this kind of change can easily become a self-reinforcing process, with the end result of which may be a totalitarian autocratic system, while a stably established classical autocracy usually has internal control mechanisms that can compensate for even a small deviation toward democracy. For human societies, the transition from democracy to autocracy is a non-coercive, naturally happening, practically self-generating process through the internal functioning of the democratic system, while the transition from a stably established autocracy to democracy is usually possible only through social coercion directly enforced by society, which typically involves turbulence in the entire community, and, as a result, typically does not create democracy as an unstable equilibrium state of social order, even in the case of the actual fall of the prevailing autocracy.
The two opposing social governance arrangements, however, have distinctly different characteristics in terms of social viability considering changes as well. Democracy typically has self-regulating mechanisms that arise from the nature of the system and are enforced by itself in response to environmental changes, which even if do not necessarily ensure adequate responses to changes, provide opportunities for correction, i.e. a democracy is a state of stability in the effects of external changes. Autocracy does not have self-regulating mechanisms for environmental changes arising from the nature of the system and enforced by itself, social responses to environmental changes are determined by the intention and intelligence of the elite governing society, which may be, or can be appropriate for society, but since there is no strict and necessary correlation between social benefit and the interests of the elite, autocracy makes the possibility of successful functioning of society not determined, i.e. autocracy is an unstable social state in terms of external changes.
Democracy as a social order is both a fragile, unstable state of a society, and also a necessary condition for societies’ self-sustaining existence in the changing environment. Artificial intelligence, with its unlimited cognitive ability, can be an effective utility to support the functioning of society, but it is also a dangerous tool in the hands of the power that governs society. Any community that cannot fully control its own governance, and the government can use artificial intelligence to perform community related tasks, or can use artificial intelligence to manage the data when accessing the community's information, the community is most likely at a great danger of existence, and putting itself at risk of suicide, and if the leadership of a society is already using artificial intelligence to control information at the community level, that community will almost certainly drift towards to authoritarian regime, i.e. it functions as a potentially lost society.
No comments