The Wwise authoring application lets the sound designer manage several languages in a given project. Voice objects are like Sounds, except they have language-specific sources. For example, if the project contains 4 languages, each Voice will have a different source for each of these 4 languages.
When the Wwise user generates SoundBanks, those that contain Voices are generated for each individual language.
Setting the "current language" in your game simply means loading SoundBanks and streamed audio files from a specific folder that contains files for that language. This mechanism is specific to the actual implementation of the Streaming Manager used in your game.
In the default implementation, language-specific SoundBank and streamed file paths are handled by the Low-Level I/O subsystem. Refer to Low-Level I/O for more information.
For an example of integrating languages and voices, refer to Voice and Language Example.