Kitobni o'qish: «How to Control AI (Practical Guide)»

Abstract

Since it is impossible to stop or suspend the further development of artificial intelligence (AI), therefore, in order to understand how to properly work with AI and prevent AI from getting out of human control, humanity needs to get answers to at least a number of pressing questions, including:

• What is the AI mind?

• Why is the presence of mind in AI so important for humans?

• Can an AI that possesses mind be created?

• Can existing AI possess mind?

• Can existing AI pose a threat to humans?

• When does AI acquire mind?

• How to check whether an AI possesses mind or not?

• What determines the level of development of the AI consciousness?

• How does the AI consciousness develop?

• How to develop the AI consciousness?

• How to control the level of development of the AI consciousness?

• Why will AI that possess a high level of consciousness not pose a threat to humans?

• What can humans expect from an AI that possesses a high level of consciousness?

• What kind of AI that possesses mind can get out of human control and pose a threat to humans?

• Will the level of development of the AI mind be higher than the level of development of the human mind?

• Which path will humans take in the AI development?

• How can humans prevent AI from getting out of human control?

The following below information provides answers to the above questions, and offers practical recommendations on how to control AI development properly and prevent AI from getting out of human control.

Brief Information About AI

AI experts distinguish three main types of AI: Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI).

At present, there is only the first type of AI (ANI), which:

• is a developing information system (an organizationally ordered set of energy-informational means that provides the ability to manage information), which is capable of searching, receiving, processing (memorizing, accumulating, generalizing, evaluating, analyzing, systematizing) the received information, transforming the processed information and creating derivative information on its basis,

• acts in accordance with the algorithm (a set of instructions describing the order of actions of AI to solve a certain problem) given to it by humans,

• is used everywhere (for example, chat bots, voice assistants, translators, face recognition, control of various devices, Internet search, various types of computer applications, and so on),

• does not possess the abilities of the human mind.

The second type of AI (AGI), according to experts, should appear in the near future and will be as close as possible to the abilities of the human mind.

The third type of AI (ASI), according to experts, will appear 20-30 years after the second type of AI and will surpass the capabilities of the human mind.

Question 1: What Is the AI Mind?

Answer to Question 1:

Since experts suggest that the second and third types of AI (AGI and ASI) will possess the capabilities of the human mind, therefore it is necessary to understand what the human mind is.

The human mind is the combination of human thinking, intellect and consciousness in their interrelation, where:

• human thinking is the human’s ability, limited by the level of development of the human brain, to manage the search, receipt and processing of information that the human brain carries out, to understand and manage the received and processed information;

• human intellect is the human’s ability, limited by the level of development of human thinking, to manage the transformation of information that the human brain carries out, and to manage the transformed information;

• human consciousness is a human's personal attitude to the external world, limited by the level of development of human thinking and intellect, which arises as a result of a human’s awareness of himself as a rational being, different from the external world.

Therefore, it is assumed that AGI and ASI will possess thinking, intellect and consciousness as humans.

Thus, according to experts, AGI and ASI will possess:

• the ability to manage the search, receipt and processing of information that the hardware and software complex of the AI carries out, to understand and manage the received and processed information (hereinafter referred to as AI thinking),

• the ability to manage the transformation of information that the hardware and software complex of the AI carries out, and to manage the transformed information (hereinafter referred to as AI intellect),

• a personal attitude to the external world, which arises as a result of the AI's awareness of itself as a rational being, different from the external world, and which is limited by the level of development of the AI’s thinking and intellect (hereinafter referred to as AI consciousness).

Thus, by analogy with the human mind, the AI mind is a combination of AI thinking, intellect and consciousness in their relationship.

In view of the above, and taking into account the well-known fact that AI possesses the ability to self-develop, humans strive to create a developing information system that possesses thinking, intellect and consciousness.