Adaptive Zero Trust Governance Framework for Autonomous and Agentic Digital Systems Using AI-Driven Risk Scoring

his preprint proposes an adaptive AI-driven Zero Trust governance framework for securing autonomous and agentic digital systems operating across modern interconnected digital infrastructures. The study integrates machine learning-based threat detection, behavioural anomaly analysis, adaptive risk scoring, and governance-oriented policy orchestration into a unified cybersecurity architecture designed for autonomous environments. The research adopts a mixed-methods approach combining quantitative evaluation using benchmark intrusion detection datasets including CICIDS2017 and UNSW-NB15 with qualitative governance analysis aligned with NIST SP 800-207 Zero Trust Architecture, NIST Cybersecurity Framework 2.0, ENISA threat intelligence guidance, and emerging AI governance principles. The proposed framework extends Zero Trust principles beyond conventional enterprise environments into autonomous AI-enabled ecosystems involving machine-to-machine interaction, intelligent automation, cloud-native systems, and adaptive digital infrastructures. The paper contributes to cybersecurity governance research by integrating continuous verification, contextual trust evaluation, dynamic risk scoring, and governance-based security decision-making into a scalable resilience-oriented architecture suitable for intelligent systems, critical infrastructure, smart environments, and future AI-driven operational ecosystems. Keywords: Zero Trust Architecture, AI-Driven Cybersecurity, Autonomous Systems, Agentic AI, Cybersecurity Governance, Adaptive Risk Scoring, Operational Trust, Intelligent Systems Security.