Artificial intelligence (AI) has become commonplace today and its many relevant technical implementations are known. Studies on the subject, on the other hand, unjustifiably overestimate the future potential of technology when it comes to the takeover of human-shaped robots based on artificial intelligence and the replacement of humans. This is a very erroneous assumption that considers the possibilities of artificial intelligence to be limitless.
To explain this, let’s first review the fundamentals of technology, starting with the question of what intelligence is. The common essence of the different formulations of dictionaries and encyclopedias is this: the ability of intelligence to acquire new knowledge, interpret new circumstances, and adapt to them. This is the general definition, including certain categories of intelligence. So there is logical, spatial, musical, and linguistic intelligence, and although it seems quite different, it is even physical intelligence. They are all genetically based, although the first four are mental while the last is physical. These categories of intelligence are well illustrated by the abstractions of mathematicians, the structures of engineers, the works of musicians and writers, and the extraordinary abilities of athletes, neurosurgeons, and fighter pilots.
The next question is what constitutes artificial intelligence. There is also a relatively accepted general definition: Artificial intelligence is the imitation and (perhaps better and faster than us) of a human activity by electronic systems such as computers and other mechanisms. Such systems are medical and social applications of image recognition software or robots for manufacturing and operating (implementing physical intelligence). The machine learning area is also included here, where the artificial intelligence software is able to make certain algorithm modifications based on the processed data. One branch of this is the artificial neural network technique, which electronically mimics the cerebral nervous system.
This paper is not intended to obscure the promise of artificial intelligence that the author of these lines also believes. On the contrary, highlighting the frontiers of technology would be in the interest of its possible future development, because indeed there is a very strong limit to artificial intelligence. The neural network can only recognize shapes from a pixel that appear to be randomly distributed than those in the comparison database. Similarly, the movements, forms, and activities of intelligent robots are electromechanical implementations of the imagination of designers. Engineering, mathematical software can optimize the wing profile of an aircraft in terms of efficiency, but they cannot invent the concept of wing profile and buoyancy.
No matter how advanced an artificial intelligence device is, it only works in the field man programmed or built into the electronics. The intellectual activity that developers used to invent an artificial intelligence application cannot be programmed. Intelligence and intelligence are completely two separate concepts. Even dictionaries make the distinction: the ability of intellectuality to think independently and to form abstract concepts. Thus, there is no artificial intelligence application that can perform an intellectual activity.
Yes, there are already robots in human form (even more so, female faces) that are “talking”. These conversations are, of course, just answers to the questions asked based on a statistical analysis of terms pre-embedded in a database. Another possible application of artificial intelligence is the construction of self-replicating robots for space travel with 3D printing technology, which was originally proposed by our compatriot János Neumann or 70 years ago. Indeed, a specially designed 3D printed robot would be able to print another version that would be able to print another version… ad infinitum. But no 3D printer can figure out a next application for machines.
Ultimately, then, the limit of artificial intelligence is a lack of intelligence. This is insurmountable to the best of our knowledge today. Perhaps in the future, when humanity finds the programming language of the intellect in the world’s most advanced bioelectronic computer, the human brain, we will really be able to develop artificial beings who think and replace us. Until then, however, an artificial intelligence tool is only as intelligent as the intelligence of the mathematician, engineer, and programmer who designed and developed the application.
Hardware, software, tests, curiosities, and colorful news from the IT world click here!