Securing IoT Devices in the Quantum Era: AI-Based Detection with Hybrid Post-Quantum Encryption

Abstract
With the advancement of quantum computing, the long-term security of IoT systems is increasingly challenged, particularly due to the vulnerability of traditional public-key cryptography. This study integrates artificial intelligence (AI) with a hybrid strategy that involves Post-Quantum Cryptography (PQC) and Quantum Key Distribution (QKD) to address this challenge. Incoming traffic from the TON-IoT dataset is processed by AI using XGBoost, Random Forest, and One-Class SVM as anomaly-detection filters. Only traffic classified as normal is forwarded to the encryption phase. Postquantum key establishment is performed using Kyber-512, and the resulting shared secret is combined with simulated BB84based QKD key material via a key-derivation function (KDF) to derive a 256-bit session key, which is then used with AES-256GCM for authenticated IoT data encryption. Experimental results demonstrate that the hybrid algorithm, along with the AI, can effectively improve IoT security. Upon optimization of the parameter , the Delivery Purity is 0.92, the Data Delivery rate is 0.80, and the Secure Delivery score is 0.94 compared to Delivery Purity of 0.77, Delivery Rate of 1.0, and Secure Delivery Fβ score of 0.87, respectively, without AI.

Author
Zina Balani

DOI
https://ieeexplore.ieee.org/document/11542476

ISSN
979-8-3315-6184-0

Publish Date: 2026-06-05