AI-Powered Privacy Protection Techniques for Smartphones
Arpitha Vasudev1, Hemanth Kumar M2, Chethan P3, Bhanushree C M4 , Anjana Pranati M5
1Arpitha Vasudev, Computer Science & Engineering, Dayananda Sagar Academy of Technology & Management
2Hemanth Kumar M, Computer Science & Engineering, Dayananda Sagar Academy of Technology & Management
3Chethan P, Computer Science & Engineering, Dayananda Sagar Academy of Technology & Management
4Bhanushree C M , Computer Science & Engineering, Dayananda Sagar Academy of Technology & Management
5Anjana Pranati M, Computer Science & Engineering, Dayananda Sagar Academy of Technology & Management
Abstract - In today’s hyper-connected digital environment, smartphone applications frequently collect and share user data without explicit consent, posing serious privacy risks. The rise of cross-app tracking, location monitoring, and behavioral profiling by advertising and analytics platforms demands proactive protection. This paper presents an AI-powered Android application that safeguards user privacy in real time. Utilizing a local VPN service, the application intercepts outgoing network traffic and employs an on-device machine learning model built with TensorFlow Lite to detect and block potential trackers and unauthorized data transmissions. It alerts users whenever sensitive resources such as the microphone, location, or camera are accessed. The app also includes a user-friendly dashboard for reviewing historical tracking attempts. Designed to be lightweight and fully functional offline, the system ensures robust privacy without relying on cloud-based monitoring. This approach offers a secure and practical solution to modern digital privacy challenges, particularly on mobile platforms.
Key Words: Android privacy, on-device machine learning, mobile security, VPN interception, TensorFlow Lite, tracker detection.