02-07-2013, 02:52 PM
Human-computer super-intelligence
Human-computer.doc (Size: 266 KB / Downloads: 17)
Abstract
The concept of human intelligence includes several types of intellectual activity – rational thinking, emotional thinking, unconscious thinking, intuitive thinking, and automatic control of the biological systems. Rational thinking refers to the cognition of the phenomena and laws of the animate and the inanimate nature. It is realized through the processing of mostly visual images and is therefore limited to the low-factor (not more than three-factor) laws of nature. Here belong, for instance, the laws of physics cognized by the mankind. All other types of human thinking are multiple- factor. The concept of artificial intelligence refers to computer simulation of the human low-factor rational thinking. This is why artificial intelligence cannot serve as the basis for ‘computer super-intelligence’. Contrary to computer super-intelligence, the ‘human-computer super-intelligence’ can solve these tasks. Human-computer super-intelligence will allow developing super-knowledge on significant multiple- factor processes – medical-biological, scientific and technical, financial and economic, organizational, etc.
INTRODUCTION
Ever since computers first appeared, people naturally wanted to turn them into intellectual assistants and partners of people. This is how the problem of artificial intelligence arose.
In order to turn computers into such intelligent devices, people decided first to ‘teach’ them (just like people) to talk, understand human speech, process visual images, position themselves in space, move, etc., i.e., basically those skills children naturally learn in their first year. And only later on it was expected to ‘teach’ computers to solve the most complicated, the so called intellectually demanding, tasks.
But what a child successfully learns in their first year, computer engineers failed to teach computers in almost fifty years. Therefore, a question arises as to what was wrong in the initial problem definition, and what the correct way to set the problem is.
ARTIFICIAL INTELLIGENCE
Soon after computers were created due to the work of many prominent scientists and engineers (Rojas, 2001), people realized the need of using them to solve more intellectual tasks than mathematical calculations (be it even the most complicated ones).
Back in 1950, the article (Turing, 1950) was published; it was afterwards reprinted many times, including under the title ‘Can the Machine Think?’ This article was the starting point for the problem of artificial intelligence. However, the term ‘artificial intelligence’ (AI) was offered later, in 1956 (McCarthy, 1959). The problem of artificial intelligence was very popular and extensively studied by electronic engineers, software developers, mathematicians, neurophysiologists, philosophers, psychologists, linguists, and others. Naturally, their interpretations of the problem, means of solving it and objectives of research differed greatly.
HUMAN-COMPUTER SUPER-INTELLIGENCE
The concerns mentioned in the previous section are raised by:
• The fact that computer super-intelligence is very likely to develop;
• Absence of any specific and persuasive positive scenarios of relationship development between the computer and human civilizations;
• Absence of any guarantees of further existence of the human civilization after the emergence of computer super-intelligence.
However, the issue of the greatest concern is that basically we are dealing with complete misunderstanding of the computer super-intelligence phenomenon and its frightening differences from human intelligence.
TECHNOLOGICAL SINGULARITIYT
However, despite huge difficulties in practical implementation of GOFAI, there is one more opinion on
the computer development prospects – it is suggested that rapid improvement of computers will lead to the emergence of computer civilization on our planet (Moravec, 1988; 1998; Vinge, 1993; Kurzweil, 2005). This scenario of AI development came to be known as
‘technological singularity’, which is defined as hypothetical explosive acceleration of the scientific and technical progress having unpredictable results.
CONCLUSION
Human rational thinking serves to quickly solve tactical aspects of the problem of survival, and multiple- factor subsystems of human thinking solve the most important strategic aspects of this problem.
6.2. Presumably, some ways of improving human intelligence which have not yet been demanded by the evolution – the use of developed intuition, conscious use of the unconscious multiple-factor thinking, etc. – can be realized, including the use of the new hardware and software.
Human intelligence which developed during millions of years of biological evolution is the most perfect system in terms of its performance, and allows creating super-intelligence on its base.