The Private Aggregation of Teacher Ensembles (PATE) framework provides a robust solution for safeguarding data privacy by leveraging multiple teacher models to generate labels for student samples. However, a significant limitation of PATE is that it does not protect the data exchanged between the student and teacher models, leaving this communication vulnerable to potential exposure. To address this gap, the paper introduces SMC-PATE, a novel approach that incorporates Secure Multi-Party Computation (SMC) into the PATE framework. This integration ensures that the interactions between the student and teacher models-specifically the process of labeling-remain confidential. In this enhanced model, teacher models operate on encrypted, unlabeled data, and the resulting labels are securely transmitted through SMC. Experimental results on the MNIST dataset demonstrate that SMC-PATE not only preserves the accuracy of the model but also enhances data privacy, making it a practical and effective solution for privacy-preserving machine learning applications.