As society is becoming increasingly digital in its interactions, the issue of data privacy is becoming more and more severe and of public concern too. Data or information can be personal, non personal or sensitive personal information of an individual. The organizations that are collecting and storing such data have the responsibility of ensuring protection to this data. They must separate the personal data and protect it which is a time consuming process. The literature reviewed for this explores a number of machine learning techniques used to classify the information into three classes, namely, non-personal, personal and sensitive personal information. This paper express that a machine learning model trained using a larger and diverse set of documents can yield a most accurate result.