Artificial Intelligence-Based Cybercrime

Abstract

Advances in the field of artificial intelligence (AI) in recent years have led to exciting capabilities, but they have also raised concerns about dangers. A major concern is about super-intelligent AI agents with the initiative, motivation, and the capabilities to attack humans. Although such super-AI are far out in the future, existing AI tools are already being used by cybercriminals to enhance the efficiency of their attacks. The immediate threat from AI includes three types of activities. First, AI gives attackers the ability to automate and enhance online cybercrime as we know it. Second, using AI opens new channels for cybercrime, including some reaching into physical space. A third type of attack using AI is against other AI systems, either in stealing confidential data or in corrupting the operation of AI algorithms to achieve the goals of the attacker. Information security professionals must be prepared for the immediate danger of malicious uses of current AI technologies.