Enhancing privacy policy comprehension through Privacify: A user-centric approach using advanced language models
2024-07-24
Summary
Privacify is an application designed to digest complex privacy policies and provide analyses in a human readable form. Technically speaking, it leverages a combination of text segmentation, summarization using Large Language Models (LLMs), and map-reduce technologies to facilitate analyses for single-document insights and for comprehensive cross-document analysis. Designed with a user-centric approach, Privacify features an intuitive interface that presents all relevant user privacy information in easy-to-understand language, complete with a detailed explainability component. Our team consisted of our primary investigator, Aisha Ali-Gombi; Katherine Perez, and myself, Justin Woodring. We evaluated our work with a ground truth analysis, including reliability and accuracy assessments, alongside rigorous functionality verification through ROUGE metrics, validating the system’s efficacy and performance.