Центральноукраїнський науковий вісник. Технічні науки. Випуск 5. Частина 1. - 2022
Permanent URI for this collection
Browse
Browsing Центральноукраїнський науковий вісник. Технічні науки. Випуск 5. Частина 1. - 2022 by Subject "artificial intelligence"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Ризики впровадження штучного інтелекту в комп’ютерні системи(ЦНТУ, 2022) Марченко, К. М.; Оришака, О. В.; Марченко, А. К.; Мельник, А. М.; Marchenko, K.; Oryshaka, О.; Marchenko, А.; Melnick, А.У статті розглянуті питання автоматизованої обробки інформації в комп’ютерних системах. Розглянуті вимоги, яким повинні відповідати комп’ютерні системи. Підкреслюється неможливість забезпечення абсолютної надійності алгоритмів та програмного забезпечення комп’ютерні системи, отже, повної адекватності отриманих рішень. Проаналізовані сфери використання штучного інтелекту та зроблені висновки про доцільність та безпечність його впровадження в окремих галузях з точки зору безпеки життєдіяльності людей та охорони праці. Since the absolute reliability of computer systems and the results of information processes that run in them can not be guaranteed, the task of research is to identify critical areas where such errors and failures are unacceptable. The main problems with the introduction of artificial intelligence in computer systems are the inability to predict all real situations and program the behavior of the machine adequately to them, lack of reliability and software errors. The input on which artificial intelligence is taught may be incorrect. In addition, artificial intelligence systems are influenced by the way of thinking and values of its developers, who are not always familiar with psychology, sociology and other humanities. These shortcomings during the use of artificial intelligence systems have led to many incidents, including fatal. The analysis of the sample of artificial intelligence error messages allowed us to determine which areas are critical errors, ie where the use of artificial intelligence systems is associated with significant risk. In particular, these are such areas as medicine, military affairs, transport, manufacturing, where people and robotic systems cooperate, hazardous industries, energy, social management, legal institutions and more. Currently, there is no regulatory and legal framework for the use of artificial intelligence, so its implementation is spontaneous, which leads to unpredictable results and accidents. Artificial intelligence used in critical infrastructures, in areas related to human health and life, belongs to the category of high risk. Based on the analysis and due to the impossibility of ensuring the absolute reliability of computer systems and their software, the authors do not recommend the use of artificial intelligence in areas related to safety, health and human life, especially large human teams. Devices using artificial intelligence systems should be marked with messages about its use with a clear warning about the partial reliability of the device in terms of safety and consumer responsibility for the use of such a device. The authors strongly discourage the use of artificial intelligence in responsible decision-making in areas related to the security of large groups of people.