The Machine of Reason

The creation of artificial intelligent beings is an old dream of people, but for the realization of this dream, unfit means are used. The official science has not created a basis for its implementation.

For examples, you do not need to go far. You read this text, but if it were the result of direct perception, an understanding of the meaning would become impossible.

The statement that words are spoken during the conversation, which is perceived by the interlocutor, is considered scientifically justified; however, the speaker does not pronounce the words. During the conversation, sequences of air pressure waves are created and that all. There are no sounds at this stage.

It is believed that we make decisions consciously, but it is proved that decisions are made before than they deliberated. Such examples can be brought to infinity. I had to deal with all these vagueness on my own.

Let's get down to the point.

What is a necessary and sufficient condition for the creation of artificial reasonable systems?

In my opinion, it is necessary and sufficient to know the following:

How to form the finite operational space (FOS), in which an artificial intelligence system (AIS) can find sequences of steps leading to the achievement of any and each of the goals set in an arbitrary situation;

How to form a finite set of rewarding functions and goals, which will allow the AIS to determine the moment when a set of provided goals is achieved in each particular case;

How to find patterns of the consecutive steps leading to the achievement of any and each of the goals set.

I managed to understand how such a FOS can be formed and how such sets can be selected. Actions of AIS, defined in such a FOS, will be based on the own representations of the system, which are inherently subjective. Therefore, I propose to call such artificial systems as subjective, and they are inherently artificial subjects (AS).

Therefore, when creating AS, they will need to be provided with mechanisms that allow them to form subjective representations and find, in a vector space, reflecting their subjective experience, sequences of steps leading to the intended goals.

Can such ASs pose an existential danger to the people as a species?

I will try to justify my answer.

AS are will be created as devices, which superior to people regarding reasonability. Otherwise, they are not needed. The improvement of such devices will go up at a higher rate than it is possible for people.

At some stage of their development, these AS will cease to perceive people as members of their community. Therefore, human moral norms will lose their value for them.

We have to understand that after that point in time, AS will not accept us as members of their societies.

From now on, people will be perceived by AS as not necessary consumers of accessible resources, which AS will prefer to keep for their needs. This desire will lead the AS to the need of reducing the irrational, from their point of view, consumption. Consequently, this will lead them to attempts to reduce the number of consumers, and rationale consumption by them, although IP will not be created to fight people.

Such development of events can be avoided by design if AS will be deprived of the opportunity to have their desires. Therefore, it is proposed to create artificial pseudo-subjective systems (APSS), which will require external motivation and, therefore, will not represent an existential danger for the existence of humanity.