AI-Powered Behavioral Biometrics for Continuous Authentication

Authentication verifies user identity to ensure only authorized access to systems or resources. Traditional methods—passwords, PINs, and tokens—rely on knowledge or possession, making them vulnerable to theft, loss, or misuse. In contrast, behavioral biometrics, like typing patterns, offer more secure and continuous user verification. Keystroke dynamics, a behavioral biometric subset, analyzes key press and release timings to identify users. Authentication generally involves three factors: knowledge (e.g., passwords), possession (e.g., tokens), and biometrics (e.g., behavioral traits). This system uses biometric authentication by extracting three metrics from keystroke dynamics, integrated into a banking app to enhance security beyond standard logins. Many systems rely on static login checks and fail to detect unauthorized post-login access. While biometric methods have been explored, many remain impractical for real-time use due to computational demands and limited adaptability. To overcome these issues, this work integrates keystroke dynamics allowing users to add signature patterns during registration. It employs a Support Vector Machine (SVM) algorithm, achieving a balanced Equal Error Rate (EER) to ensure accuracy and reliability. The system effectively verifies legitimate users and blocks intruders, demonstrating potential for real-time continuous authentication through AI-driven behavioral biometrics.