Peculiarities of the confrontation between defensive and offensive artificial intelligence in cyberspace

Abstract

The use of artificial intelligence in cyberspace significantly changes the course of the confrontation between defensive and offensive technologies. Thus, at the current stage of development of information technologies, artificial intelligence systems and/or models are used not only to strengthen cyber defence systems, but also to develop new types (kinds) of information-destructive impacts in the form of adaptive cyber attacks that can potentially avoid detection by existing defence systems. Cyberattacks created with the use of artificial intelligence are characterised by applied novelty, complexity, speed of adaptation and scalability, which makes existing methods of detecting cyberattacks almost ineffective, which in turn poses a serious threat to information systems of both state and commercial purposes. In addition, there has been a significant increase in the number of cases of cyberattacks and malware created using artificial intelligence models recently, as a result of their public availability and the virtual absence of restrictions on their use. The article analyses typical approaches to the training and use of both defensive and offensive artificial intelligence in cyberspace for the purpose of conducting defensive and offensive (counter-offensive) cyber operations. The author identifies their common and distinctive features, as well as interacting factors and interrelationships in the course of confrontation, which makes it possible to form the basis for solving the scientific and applied problem of preventing cyberattacks created using artificial intelligence technologies. Based on the obtained features of the confrontation between defensive and offensive artificial intelligences in cyberspace, the author suggests ways for further scientific research to ensure that the benefits of using artificial intelligence technologies for malicious purposes can be levelled.