Revolutionizing Cybersecurity: A Generative AI-Powered Malicious File and URL Detection System

Abstract

The growth of polymorphic malware and malicious URLs highlights the weaknesses of traditional signature-based and heuristic defenses, especially against zero-day threats. This research proposes an AI-driven detection framework that combines Python-based static feature extraction with OpenAI’s GPT-4.1-mini to classify threats using structured prompts, offering explanations and confidence scores. Based on the Cognitive Security framework, it shifts cybersecurity from reactive rules to adaptive, intelligence-driven protection. Initial conceptual results suggest better zero-day detection, fewer false positives, and clearer forensic insights, demonstrating the transformative potential of generative AI in cyber defense [1].