Blockchain-Integrated AI Framework for Secure IoT Communications

Abstract

Internet of Things (IoT) deployments suffer from weak device identity, insecure message transport, and limited auditability issues amplified by device heterogeneity and resource constraints. This paper proposes a Blockchain-Integrated AI Framework (BIAF) for secure IoT communications that combines (i) lightweight edge AI for near-real-time intrusion/anomaly detection, (ii) permissioned blockchain for immutable audit logs and decentralized trust, and (iii) smart-contract access control to enforce communication policies. The framework supports constrained protocols (e.g., MQTT and CoAP) and aligns with baseline IoT cybersecurity capabilities recommended by NIST. We evaluate BIAF using a benchmark methodology inspired by public IoT intrusion datasets (e.g., CICIoT2023 and BoT-IoT) and provide a reproducible synthetic dataset and metrics. Results show that blockchain-backed attestation and auditability can be achieved with manageable overhead while improving integrity, accountability, and cross-domain trust.