Problem
Research questions and friction points this paper is trying to address.
user reviews
permissions
AI apps
privacy concerns
security
Analysing User Reviews to Identify User Concerns Around Permissions in AI Apps
🤖 AI Summary
This study addresses the gap in security and privacy best practices among AI-powered mobile application developers, which often leaves users responsible for monitoring permission usage themselves. To overcome the challenge of analyzing unstructured user reviews, the authors propose a weakly supervised learning approach that leverages AI-generated permission-related comments as weak supervision signals to automatically construct a labeled training set from large volumes of unlabeled user feedback. A classification model trained on this dataset achieves 82% accuracy in identifying permission-related content in real user reviews. The findings reveal that users’ concerns are predominantly shaped by their overall sentiment toward the app rather than specific permission types. This work offers valuable insights for developers, platform providers, and users regarding privacy communication and permission design.
📝 Abstract
Artificial intelligence is increasingly embedded in everyday software, making its integration into mobile apps inevitable. However, AI mobile app developers are not always versed in security and privacy best practices, leaving users to monitor their own security and understand how apps use their data. App reviews capture real user experiences, helping others make informed decisions before downloading. This paper presents a machine learning model for classifying AI app reviews into permission-related categories. Because user reviews are unstructured, assembling a conventional labeled training set is difficult. To address this, AI-generated security and permission reviews are used to identify relevant training examples from a large corpus of human-written reviews, eliminating the need for manual annotation. The proposed approach classified permission reviews with an accuracy of 82%. Analysis shows that users organise their concerns by sentiment toward the requesting app rather than specific permission types, with implications for users, developers, and platform administrators.
Innovation
Methods, ideas, or system contributions that make the work stand out.
AI-generated reviews
permission classification
machine learning
user sentiment
privacy concerns
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
