Hidden in Plain Sight: Detecting Illicit Massage Businesses from Mobility Data
A new system analyzes location patterns to identify high-risk businesses, improving detection rates by 5.3x.
Researchers from multiple institutions developed an AI model that uses mobility data to detect illicit massage businesses (IMBs). The system analyzes temporal visitation patterns, dwell times, and visitor catchment areas, achieving 0.97 AUC. Using positive-unlabeled learning, it identifies four operational signatures of high-risk establishments. Law enforcement can use this decision-support system to prioritize inspections, with the top 10% of flagged locations containing 53% of known illicit operations.
Why It Matters
Provides law enforcement with a powerful, data-driven tool to combat human trafficking more efficiently by targeting resources effectively.