Adaptive-AI-ZeroTrust-Chain: blockchain-backed dynamic zero-trust enforcement via artificial intelligence

The proliferation of Internet of Things (IoT) devices and distributed computing environments has intensified the demand for robust, adaptive security frameworks capable of continuous trust evaluation. Traditional perimeter-based security models fail to address the dynamic nature of modern network ecosystems, where device behavior evolves continuously and adversarial threats adapt in real-time. This paper introduces Adaptive-AI-ZeroTrust-Chain (AAZTC), a novel framework that integrates artificial intelligence-driven dynamic trust boundary modeling with blockchain-based verifiable access logging to enable granular, auditable zero-trust enforcement. The proposed architecture employs deep reinforcement learning algorithms for continuous behavioral analysis and trust score computation, while leveraging smart contracts on a permissioned blockchain to ensure immutable, transparent access decision records. The framework incorporates a lightweight post-quantum cryptographic module to future-proof security against emerging quantum computing threats. Extensive experiments conducted on the NSL-KDD and CICIDS2017 datasets demonstrate that AAZTC achieves 98.73% detection accuracy, 97.89% precision, and 98.21% F1-score, outperforming state-of-the-art baseline methods by margins of 3.2–5.8%. The system maintains low latency characteristics with average trust decision times of 12.4 ms, making it suitable for real-time IoT deployments. Ablation studies confirm the synergistic contributions of each architectural component, validating the comprehensive design philosophy underlying AAZTC.