Advancing Data Privacy in REST APIs: A Comparative Study of Machine Learning Techniques for PII Detection

The increasing use of REST APIs for data exchange has heightened the risk of exposing Personally Identifiable Information (PII), necessitating the development of automated detection systems to ensure software quality and compliance with regulatory standards. Global privacy regulations, including the GDPR and Indonesia's UU PDP, require organizations to effectively safeguard PII across various data formats, particularly unstructured data typically transmitted through APIs. Conventional approaches, such as rule-based detection, are inadequate for addressing the dynamic and semi-structured nature of API responses. This paper presents the application of machine learning techniques, specifically Support Vector Machines (SVM) and Naive Bayes classifiers, for detecting PII in REST API responses. The SVM model achieved an accuracy of 80.96%, a precision of 90.11%, a recall of 80.96%, and an F1-score of 84.37%, surpassing the performance of the Naive Bayes model. The results highlight the potential benefits of incorporating machine learning-based PII detection systems into the software development lifecycle to enhance API security, software quality, and compliance with privacy regulations.