AI in Cybersecurity: A Double-Edged Sword-Case Studies on Its Defensive and Offensive Applications
NUR ALYYANA SOFEYA MOHD JEFRYY, Nurfarah Izzati Binti Arman, Nurfarah Najwa Binti Mohd Takri · 2025
The field of digital defense and offence has changed because of the incorporation of artificial intelligence (AI) into cybersecurity. Advanced features like anomaly detection, phishing prevention, and intelligent endpoint protection are made possible by AI technologies, which help businesses identify and address threats more quickly and precisely. Malicious actors, however, are rapidly using the same skills to carry out complex assaults, such as phishing produced by AI, deepfake-based fraud, and adversarial exploitation of learning models. In this paper, a case study analysis is given comparing positive and negative aspects of AI in cybersecurity. Taking the existing applications into practice like the anomaly detection system developed by Darktrace, BERT-based phishing detection, and hybrid endpoint monitoring systems and malicious applications such as deepfake impersonation and phishing attacks using LLMs, the paper presents the position of both sides of AI as protection mechanism and as a potential threat. These results reveal that being responsible for the use of AI and keeping AI a tool of strength but not against it may be seriously challenged by ethical guidelines, training against adversaries and regulation of using AI in everchanging cybersecurity scenario.