The Duality of Artificial Intelligence: From Proactive Threat Detection to Automated Cyber Attacks

Abstract

Artificial Intelligence (AI) has become one of the most controversial and, at the same time, most powerful technologies of the digital era- capable of protecting entire systems while also rendering them vulnerable. On one hand, AI is transforming cybersecurity through its ability to detect anomalies, analyze massive volumes of data in real time, and automatically respond to incidents- significantly reducing the time interval between the emergence of a threat and its neutralization. On the other hand, these same capabilities are being exploited by malicious actors to automate cyberattacks, generate sophisticated phishing campaigns, create convincing deepfakes, and develop adaptive malware that is difficult to detect through conventional methods. This contradiction places AI at the center of an unprecedent technological race, in which the boundary between protection and threat becomes increasingly fluid and difficult to define. The purpose of this research is to comparatively analyze the two dimensions of artificial intelligence use in cybersecurity- defensive and offensive- within the context of todays digital economy, with the aim of identifying mechanisms through which AI can be governed responsibly, maximizing its protective benefits while minimizing the risks associated with its malicious use. CZU: 004.056:004.8; JEL: O33