Microsoft does not share the data you submit for translation with anybody. Microsoft does not take any samples of any text submitted for translation using the Translator Text API. More information is outlined in the No-Trace webpage.
The material you use for training on the Translator Hub is stored encrypted in your workspace. You may invite whoever you like into the workspace, identified by an email address and authenticated with a Microsoft Account. The people you designate as co-owner have access to your training material and training runs the same way you have. Microsoft will not share the data with anyone else. You can delete the data in your workspace at any time, as long as it is not used in any training. Microsoft will use the data that you include in a training to build the custom translation system for you. The custom translation system will be deployed on Microsoft's servers, and are then usable via the Microsoft Translator API. Anyone who knows the category ID of your custom system can use it using their own Microsoft Translator Text API subscription account. There is no method allowing you or anyone else to generate the original training material out of the trained system. Microsoft may use the data in your workspace to improve Microsoft Translator only if the "Keep data private" box in the Settings tab on the Translator Hub is unchecked.
Copyright and Intellectual Property:
Translation, or use of a document for training on the Microsoft Translator Hub, does not alter the ownership of the intellectual property contained therein.
Last update: March 9, 2017