Proposed Hybrid Ai Multi-Factor Authentication Based On Individual Cognitive-Behavioural Pattern-Based Interaction Habits

Background: Identity attacks, such as credential theft, phishing, spoofing, and replay, are becoming increasingly prevalent, undermining the efficacy of MFA. Although knowledge-, possession-, and biometric-based techniques improve security, they can still suffer from theft, variability, and usability issues. Although behavioral biometrics hold promise, they are constrained by high surveillance costs, neuromotor variance, and catastrophic forgetting in adaptive AI systems. Methods: A neuro-inspired hybrid MFA, where the novel fourth auth factor is based on procedural memory, is proposed. This aspect captures sequences of discrete cognitive–behavioral interactions, such as the order of application launches, navigation flow, and refresh patterns. The interaction model is based on deterministic finite automata (DFA), and temporal dynamics are modeled using Gaussian mixture models (GMM). Constraint-based multi-session registration stabilizes identity learning, and checkpoint-based AI updating mitigates catastrophic forgetting while maintaining baseline performance. Results: Experiments on synthetic and real interaction datasets demonstrated high robustness against replay and spoofing attacks. The system provided the following results: an Equal Error Rate (EER) of 2.4%, a 62% reduction in computational load, and a verification latency of 280 ms. Under neuromotor variability enactment, the False Rejection Rate (FRR) was 3.1%, compared to adaptive AI genre systems, which had an FRR of 12.1%. Conclusions: The cognitive–behavioral MFA framework integrates procedural memory characteristics with available credentials to yield a robust, low-latency, neuroinspiration-based authentication model. Consequently, it overcomes the drawbacks of traditional and behavioral systems, thus proposing secure and adaptive human–AI authentication frameworks.