@Sarmad Raza Thanks for the question. The standard voice model (also called a voice font) has been trained by using Microsoft-owned data and is already deployed in the cloud. Speech services do not store customer data, the customer data is automatically deleted after processing.
For Compliance and Privacy:https://azure.microsoft.com/en-us/support/legal/cognitive-services-compliance-and-privacy/