A. A. Zhdanov – Dr.Sc. (Phys.-Math.), Professor, Chief Research Scientist, Department of Scientific Preparation of Projects, Institute of Precision Mechanics and Computer Engineering n.a. S.A. Lebedev of the Russian Academy of Sciences; Professor, Basic Department of Computer Engineering, Faculty of Radio Engineering and Cybernetics, Moscow Institute of Physics and Technology

E-mail: a.zhdanov@mail.ru

Modern advances of artificial intelligence (AI) systems generate strong enthusiasm among both developers and consumers, despite the fact that these systems are still relatively simple, and the emergence of much more powerful systems is still in sight. However, every invention has two sides of the coin – useful and harmful. While much is being said about the alleged benefits of AI systems today, there is very little about the possible harm from these systems. However, many of the dangers of the expected widespread distribution of AI systems can already be foreseen, they are very dangerous. Understanding of this problem should cause appropriate forces to prevent the expected damage from AI systems.

The aim of the work is to identify the problems that are likely to arise with the widespread use of AI systems. This should help customers, investors, developers and other parties representing the interests of consumers to provide timely means and ways to minimize the potential harm that can be caused to society by AI systems.

Several potentially dangerous properties of AI systems and the consequences of their widespread using have been listed and considered in the article. Justifications and examples of such dangers have been given.

The listed dangers of wide distribution of AI systems can give rise to their discussion by interested parties. In result of the discussion the deeper understanding of the problem can be reached.

1. Averkin A.N., Gaaze-Rapoport M.G., Pospelov D.A. Tolkovyj slovar' po iskusstvennomu intellektu. M.: Radio i svyaz'. 1992. (in Russian)
2. Zhdanov A.A. Avtonomnyj iskusstvennyj intellekt. M.: BINOM. Laboratoriya znanij. 2008. (in Russian)
3. McCarthy J. Recursive functions of symbolic expressions and their computation by machine. Part I. Communications of the ACM. 1960. V. 3. № 4. P. 184–195.
4. Leshchev S.V. Poststrukturalistskie meditatsii. M.: MIFI. 1998. (in Russian)
5. Metz C. Why the final game between AlphaGo and Lee Sedol is such a big deal for humanity. Wired News. 14 March 2016.
6. Komp'yuter Deep Blue pobedil v shakhmaty chempiona mira Garri Kasparova [Elektronnyj resurs]. URL: https://denvistorii.ru/ 11-maya/kompiuter-deep-blue-pobedil-v.html. (in Russian)
7. Bengio Y. Learning deep architectures for AI. Foundations and trends in machine learning. 2009. № 2. P. 1–127