Evaluating the Impact of Artificial Intelligence in Adaptive Access Control Models

Abstract

The growing dynamism and complexity of cybersecurity threats necessitate real-time, adaptive access control systems that cannot be constrained by traditional access control models, such as Role-Based Access Control (RBAC) and Attribute-Based Access Control (ABAC). This paper will examine how Artificial Intelligence (AI) can be incorporated into Adaptive Access Control (AAC) models to improve the effectiveness of security decision-making, system flexibility, and effectiveness. It used a mixed-methodology, i.e., it combined quantitative analysis of system logs through machine learning models such as Support Vector Machine (SVM), Decision Tree (DT), and Artificial Neural Network (ANN) with qualitative information obtained in interviews with experts and surveys. The findings show that AI-based access control has a strong capability to detect anomalies and minimize the reaction time and dynamically change permissions according to user behavior and risk context, which are beyond the scope of fixed RBAC and ABAC systems. Another key finding is that data quality, retraining of the model, and Explainable AI (XAI) are important to maintain the transparency, trust, and ethical adherence of automated decisions. Besides, the paper confirms a pragmatic possibility of Cognitive Access Control (CAC) where AI constantly learns and optimizes access decisions to proactively predict threats. These findings add to the conceptual knowledge of the AI socio-technical systems and present a roadmap to organizations aiming to install intelligent, resilient, and context-sensitive access control systems, and consider operational, ethical, and regulatory challenges in cybersecurity governance.