How Private is Your Browsing? Detecting GDPR Violations Through Robust Natural Language Processing and Thorough Static Program Analysis Techniques

In the current age of the internet, online users are becoming increasingly cautious of the use of their private information. As a protective measure, the European Union (EU) instated the General Data Protection Regulation (GDPR) to protect the privacy of its citizens. However, many websites still do not follow the regulations listed in GDPR. We present a system that can thoroughly check websites using WordPress for GDPR violations through static program analysis, while also being able to check other websites for violations in the privacy policy through our natural language processing (NLP) model. This system improves on previous findings by using the website's URL to automate the process of gathering the data needed for analysis, which we accomplish with web scraping techniques. Finally, we present our NLP model with an accuracy of 80% that utilized transformers and transfer learning. This paper also gives various methods to apply and improve the system presented in this paper.