Privacy of Groups in Dense Street Imagery

2025-05-11

AI Summary

This paper exposes a critical group privacy risk in dense street imagery (DSI): even after facial and license plate blurring, AI models can infer sensitive group attributes—such as religion, political affiliation, and sexual orientation—from high-density visual contexts, thereby exceeding the protective scope of conventional individual de-identification. Method: We introduce the first systematic typology of identifiable groups in street scenes, grounded in contextual integrity theory to define novel dimensions of group privacy; we integrate penetration testing, visual feature analysis, group behavior modeling, and a dedicated privacy assessment framework. Results: Empirical validation across 25.23 million vehicle-mounted street images from New York City demonstrates the feasibility of inferring sensitive group affiliations. Our work advances DSI governance beyond individual-level de-identification toward a new paradigm centered on proactive group privacy protection.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Privacy
• Philosophy and Ethics of AI: Privacy & Security

Application Category

• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Spatially and temporally dense street imagery (DSI) datasets have grown unbounded. In 2024, individual companies possessed around 3 trillion unique images of public streets. DSI data streams are only set to grow as companies like Lyft and Waymo use DSI to train autonomous vehicle algorithms and analyze collisions. Academic researchers leverage DSI to explore novel approaches to urban analysis. Despite good-faith efforts by DSI providers to protect individual privacy through blurring faces and license plates, these measures fail to address broader privacy concerns. In this work, we find that increased data density and advancements in artificial intelligence enable harmful group membership inferences from supposedly anonymized data. We perform a penetration test to demonstrate how easily sensitive group affiliations can be inferred from obfuscated pedestrians in 25,232,608 dashcam images taken in New York City. We develop a typology of identifiable groups within DSI and analyze privacy implications through the lens of contextual integrity. Finally, we discuss actionable recommendations for researchers working with data from DSI providers.

Problem

• Addressing group privacy risks in dense street imagery datasets
• Demonstrating harmful group inferences from anonymized dashcam images
• Analyzing privacy implications for identifiable groups in urban data

Innovation

• Penetration test on 25M dashcam images
• Typology of identifiable groups in DSI
• Privacy analysis via contextual integrity