SOME ASPECTS OF APPLICATION OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES IN INFORMATION SECURITY (REVIEW)

Abstract

Artificial intelligence (AI) technologies are one of the most dynamically developing areas of information
processing. AI technologies are used both to ensure the information security and to organize attacks
on information security tools. AI systems themselves may contain vulnerabilities and be susceptible
to various types of attacks. The article analyzes some aspects of the use of AI technologies in information
security tasks. Within the framework of the task of biometric identification, threats of falsification of biometric
identification characteristics in order to obtain access rights, and ways to counter such threats are
considered. The advantages of using AI in protecting information in computer systems and networks in
comparison with traditional means of protection are analyzed. Using the example of an acoustic channel
of information leakage from a keyboard, the use of AI technologies for processing data from technical
leakage channels is illustrated. Methods for increasing the information content of such channels using temporary convolutional networks and image classification models, as well as ways to counter them, are
considered. Special attention is paid to information security issues in increasingly popular systems for
compressing and transmitting information without significant semantic losses (Semantic Communications).
A number of information security issues that arise when using large language models such as
ChatGPT, capable of massively generating unique “human-like” content and using it to organize phishing
and other social engineering attacks, are considered. An attack on AI systems using a covert channel is
described. Attention is paid to the need to develop trusted artificial intelligence technologies