Ensuring Privacy in Natural Language Processing: A PRISMA Oriented Literature Review

Using Natural Language Processing (NLP) as a discipline of machine learning, organizations can better organize their data in order to better represent their internal knowledge. To make NLP models easier to re-use in other contexts, they should be protected accordingly. This raises the question of the appropriate privacy-ensuring technology (PET). To be able to address this question this paper conducts a literature review regarding the ensuring of privacy in NLP following the PRISMA framework. After following the identification process of significant sources, 22 valuable studies were selected. Some of these studies have shown promising results, however the field of privacy ensuring in NLP is still uncategorized with different approaches difficult to compare.