Abstract: Cloud computing, remote work, the Internet of Things (IoT), and internationally distributed network environments are all fast changing, making traditional perimeter-based security solutions useless against emerging cyberthreats. Zero-Trust Architecture (ZTA) is based on the principle that "never trust, always verify." Access, authorization, and authentication are continuously needed even though the least amount of electricity is used. Due to their reliance on manual installation, strict limitations, and laws that forbid modifications, many of today's Zero-Trust systems find it difficult to adjust to changing attack patterns and sophisticated user behavior.

The Zero-Trust Architecture proposed in this work makes use of artificial intelligence to facilitate dynamic trust evaluation and real-time access limitation. They do this by using machine learning and advanced analytics. Real-time risk and trust scores for apps, devices, and users are computed using contextual awareness, behavioral analytics, and continuous monitoring. The system employs deep learning-based behavioral modelling, AI-driven anomaly detection, and reinforcement learning for policy optimization to identify who has access to what and what threats are likely to materialize.

A thorough architectural framework is presented in the article, which includes important elements like the trust evaluation module, the AI-powered policy engine, the telemetry gathering layer, and the policy implementation locations. In tests using real-world cybersecurity datasets, our approach fared better than current rule-based Zero-Trust systems in terms of response time, false positives, and threat detection. According to the results, AI can change Zero-Trust from a static security framework into an active defence system that adjusts itself. This study contributes to the expanding literature on intelligent cybersecurity by offering a Zero-Trust framework that is future-proof, scalable, and dependable. Because of this, it may be applied to both enterprise and contemporary cloud-native systems.

Keywords: Machine learning, AI, cloud security, network security, continuous authentication, anomaly detection, trust assessment, adaptive access control, and zero-trust architecture are a few examples. Machine learning, AI, cloud security, network security, continuous authentication, anomaly detection, trust assessment, adaptive access control, and zero-trust architecture are a few examples.

Downloads: | DOI: 10.17148/IJIREEICE.2025.131221