Aplicación de la inteligencia artificial en la detección de amenazas de phishing en redes de área local

Abstract

The security of computer systems in local area networks (LANs) is facing an increasing number of challenges as a result of the growing sophistication of cyber threats. This study addresses the issue of the inadequacy of traditional intrusion detection systems in the context of advanced attacks, proposing an innovative solution based on the use of artificial intelligence (AI). The overarching objective of this research is to develop a machine learning-based mechanism for the detection of phishing threats in local networks. This mechanism has been implemented and evaluated in a simulated environment with GNS3. The methodology employed, CRISP-DM, provided a framework for the process, from initial data understanding to model validation. The primary findings indicate that the AI-based detection system markedly outperforms traditional methodologies, exhibiting high accuracy and precision in threat identification, with evaluation metrics indicating greater than 95% accuracy. This breakthrough not only enhances security and data integrity in LANs, but also establishes a precedent for future cybersecurity research. This work has significant implications at multiple levels. From a scientific perspective, it contributes to the advancement of advanced threat detection techniques. From a technological standpoint, it offers a reliable and adaptable solution to protect critical infrastructures. From an economic perspective, it reduces the costs associated with security incidents. In conclusion, this research not only proposes an effective solution to a current problem, but also opens new avenues for the implementation of AI in cyber protection, consolidating its theoretical and practical relevance in the field of computer security.