New Machine Learning Framework to Safeguard Consumers’ Mobile Location Data

Academics from Carnegie Mellon University, the University of Virginia, and New York University have developed a machine learning-based framework that preserves the usefulness of mobile location data for advertisers while reducing privacy risks for consumers. The researchers created an obfuscation scheme that suppresses certain visited locations based on the individual’s risk level to protect privacy and maintain data utility for advertisers. Their framework accommodates various types of risks and utilities while respecting acceptable levels set by both consumers and advertisers. Through testing, the researchers demonstrated that this framework outperformed previous methods. Location data from smartphones generates mobile location data that is useful for personalized recommendations, location-based advertising, and other applications. However, sharing location data can expose individuals’ personally identifiable information, leading to privacy risks. This study offers a crucial tool to ensure a secure and privacy-aware environment in the multi-billion-dollar location data ecosystem.