Mohammad Baradaran
Volume 1, Issue 1
Date of Publication: 11 December 2025
The paradigm of digital trust within the domain of electronic banking is undergoing a substantive shift from static, credential-based security protocols to dynamic, behavior-based assurance mechanisms. Traditional models, which are predicated upon discrete authentication events, demonstrably fail to provide continuous and adaptive security, a deficiency that results in both user friction and significant vulnerabilities to post-authentication threats. The present study propounds a novel framework for a Dynamic Digital Trust Model, achieved through the synergistic integration of Artificial Intelligence (AI) for behavioral analytics with Blockchain technology for immutable verification and governance. The proposed architecture establishes a "Dynamic Trust Score" (DTS) for each user, which is a quantifiable metric designed to evolve in real-time in accordance with their multi-faceted behavioral patterns. The core of the framework comprises: (1) a Long Short-Term Memory (LSTM) based autoencoder model, which continuously analyzes high-frequency user interaction data (e.g., session timing, navigation patterns, transaction rhythms, and keystroke dynamics) to detect anomalies against an established behavioral baseline; and (2) a permissioned Hyperledger Fabric blockchain that serves as a tamper-proof "Trust Ledger," for the purpose of immutably recording the DTS and the cryptographic hashes of critical, trust-defining actions. The framework was subjected to validation using a comprehensive, synthetically generated dataset simulating e-banking user behaviors, including sophisticated impersonation and account takeover attempts. The results indicate exceptional performance in the differentiation of legitimate user sessions from anomalous ones, achieving an Area Under the Curve (AUC) of 0.98 and an F1-Score of 0.97. This research demonstrates that the fusion of AI-driven behavioral biometrics and blockchain-based immutability engenders a resilient, transparent, and adaptive trust ecosystem, thereby significantly enhancing security while enabling a truly frictionless and risk-adjusted user experience.
Digital Trust, Electronic Banking, Blockchain, Artificial Intelligence, Behavioral Analytics, Long Short-Term Memory (LSTM), Dynamic Trust Score, Cybersecurity, Continuous Authentication
Mohammad Baradaran, Independent Researcher, Iran.
Baradaran, M. (2025). A Framework for a Dynamic Digital Trust Model in E-Banking using a Synergistic AI and Blockchain Approach to Customer Behavioral Analytics. J Cogn Comput Ext Realities, 1(1), 01-19.