Enhancing Password Security: Leveraging AI for Typing Pattern Authentication

With the increasing need for improved cybersecurity, the limitations of traditional password-based authentication systems have become more evident. This study proposes an AI-enhanced authentication method using keystroke dynamics to reinforce password security. By leveraging machine learning algorithms to analyze users' unique typing patterns, the system offers an added layer of defense against unauthorized access, particularly in mitigating threats such as credential theft, phishing, and brute-force attacks. The research explores various machine learning models, including Logistic Regression, Random Forest, and Support Vector Machines (SVM), to determine their effectiveness in classifying genuine users versus impostors. Experimental results show that Random Forest achieved an accuracy of 93.2%, making it the best-performing model in this study. This high accuracy demonstrates the reliability of keystroke dynamics as a method for continuous user authentication. The system’s feasibility is further underscored by its ability to scale easily within existing infrastructures without significant hardware changes. Furthermore, the solution has potential for widespread integration into real-world security systems, providing a robust and non-intrusive method for user verification across various platforms, from personal devices to enterprise-level applications.