Language is always situated, that means it is uttered in the situation at a place, time or date.
“In spoken language analysis, an utterance is the smallest unit of speech.”
(https://en.wikipedia.org/wiki/Utterance, 17/10/17)
It also depends on the individual speaker and how the speakers’ mood is. This factor leaves an imprint on the utterance in the speech. The voice which we use in the natural language processing has the spoken word and the messages in the spoken word. It also has all this information about the speaker.
For example, you could say something like “Could you help me?” in so many different forms, that someone could react differently to the questions, then they would do it when it would be pronounced in another way. So the pronunciations play a big role in this topic, out of the pronunciations you could hear, if someone is happy or if someone is really annoyed by something. With the Natural-Language-Processing you could get more than the questions the user asked for, you could get a wider variety of information from the user.
With the additional information, the program could react differently, when you’re stressed out the program could give you only the main information that you are looking for. If the Natural-Language-Processing recognizes that you feel sad, it could ask you if you want to call your best friend.
The possibilities of what companies could do with this information cannot be overlooked.
The conclusion is, yes Natural-Language-Processing needs an ethic discussion.