AI-powered Phishing Detection Systems: A Comparative Study

Phishing attacks remain one of the most persistent threats in cybersecurity, evolving in complexity and sophistication.Traditional detection mechanisms, often reliant on static rules and blacklists, struggle to keep pace with these dynamic threats.This research paper presents a comparative study of AI-powered phishing detection systems, examining the performance of various machine learning and deep learning models.Utilizing datasets containing real-world phishing and legitimate samples, models such as Decision Trees, Random Forests, Support Vector Machines, Convolutional Neural Networks, and Long Short-Term Memory networks are evaluated based on accuracy, precision, recall, and ROC-AUC scores.Our results demonstrate that deep learning approaches, particularly sequence-based models like LSTM, achieve superior performance by capturing intricate patterns in URLs and textual content.Furthermore, the study highlights the importance of feature engineering and data diversity in building resilient detection systems.By providing a comprehensive performance analysis, this study aims to assist researchers and cybersecurity practitioners in selecting appropriate AI models for phishing detection tasks.The findings also pave the way for future research into hybrid models and real-time, adaptive phishing defences.