Human Behavior as a Key: An Innovative Fusion of AI and MFA

Strong authentication methods are now essential due to the growing dependence on digital platforms. Static passwords and other conventional techniques are inadequate because they are susceptible to phishing and brute force attacks. Despite adding a physical authentication component to improve security, USB token-based multi-factor authentication (MFA) systems are still vulnerable to compromise when both the password and the token are stolen. This paper suggests a sophisticated authentication framework that combines behavioral anomaly detection powered by artificial intelligence with USB token-based protocols. By examining user-specific behavioral patterns, such as keystroke dynamics, mouse movements, and login session characteristics, the suggested system adds an adaptive, context-aware security layer. Experiments on publicly accessible keystroke datasets show anomaly detection accuracy of over 90%, highlighting the approach's viability and resilience. The foundation for creating intelligent authentication systems that can withstand complex cyberthreats is laid by this work.