AI-DRIVEN PRIVACY AND SECURITY FRAMEWORKS FOR SMART HOMES: A COMPREHENSIVE SURVEY

Abstract

Smart homes have major evolution in residential automation, which integrates Internet of Things (IoT) technologies to improve comfort, energy efficiency, and remote accessibility. Widespread adoption of smart home systems has faced critical privacy and security challenges due to uses of heterogeneous communication protocols, constrained computational power of sensor nodes, and lack of standardized encryption authentication mechanisms and strong AI based attack detection and mitigation system. This paper presents a comprehensive survey of privacy and security issues in IoT-enabled smart homes, emphasizing on vulnerabilities at the different layers of IoT communication system. It discussed major communication protocols ZigBee, Bluetooth, Wi-Fi, Z-Wave, 6LoWPAN, LoRaWAN, and NFC and highlighting their operational weaknesses and potential attack vectors such as denial-of-service, spoofing, replay, and man-in-the-middle attacks. A systematic research methodology was applied to analyze 56 key publications from 2014 to 2025 drawn from IEEE, Scopus, Springer, arxiv journal and MDPI databases. The reviewed studies reveal a growing application of Artificial Intelligence (AI) and Machine Learning (ML) methods for intrusion detection, anomaly recognition, and adaptive security enhancement. Findings indicate a clear transition from traditional cryptographic protection toward intelligent, data-driven, and decentralized security frameworks. The study identifies major research gaps related to data integrity, edge level AI based attack detection system implementation, federated learning, and explainable AI for security and privacy improvement IOT ecosystem. ML techniques (DT, RF, ANN, CNN, and LSTM,) demonstrated over 90% accuracy in attack detection, yet issues such as protocol vulnerabilities, data privacy, and limited computation capabilities remain. The review highlights the need for lightweight, explainable, and privacy-preserving IDS solutions to strengthen security in resource-constrained smart home environments.