Artificial Intelligence-Based Zero Trust Security Approach for Consumer Industry

Development in Internet technology made consumer electronics growth to another extent where several consumers from all over the world utilize various essentials through recent development. However, consumer electronics based devices could be vulnerable to cyber attacks if it is not appropriately secured. In this research work, we proposed AI-enabled deep learning model based zero trust security (AIDL-XTS) framework for verification and authentication for devices, users, and applications for every access request. We use smartphone sensor data for user authentication using Deep CNN-BiLSTM network. Furthermore, we proposed Bayes theorem based trust score to evaluate the zero trust security. This proposed framework assumes all users, devices, and applications are un-trusted which requires verification and authentication for every access request, regardless of the user’s location or device. To evaluate the trust score in the zero trust security model, Bayes theorem-based trust score (Bayes-TSC) model is proposed. The performance of model is analyzed over three datasets: WISDM-HARB, HMOG, and UCI-HAR, using four metric measures: accuracy, equal error rate, success rate, and authentication time. From the results, the performance of proposed framework outperforms when compared with traditional benchmark deep learning models for user authentication while protecting against unauthorized access in minimal authentication time.