Digital Divides in Scene Recognition: Uncovering Socioeconomic Biases in Deep Learning Systems

2024-01-23

arXiv.org

AI Summary

This paper reveals significant socioeconomic bias in deep convolutional neural networks (dCNNs) for scene recognition: models exhibit lower accuracy and confidence on residential images from low-socioeconomic-status (SES) regions and disproportionately assign discriminatory labels (e.g., “ruin”, “slum”). Method: Leveraging nearly one million geotagged residential images globally and across the U.S., we integrate public socioeconomic statistics—including income and Human Development Index (HDI)—with spatial metadata to construct a multi-source framework for quantifying bias. Contribution/Results: We provide the first empirical evidence of systematic socioeconomic bias in computer vision models across geographically and racially diverse scenes. We introduce a novel “sociostatistical–model attribution” analytical paradigm that jointly interprets demographic data and model behavior. Results are robustly replicated across multiple international and U.S.-specific benchmarks, offering critical insights and methodological foundations for fair AI design—particularly in high-stakes applications such as automated property valuation and intelligent surveillance systems.

Technology Category

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

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Computer-based scene understanding has influenced fields ranging from urban planning to autonomous vehicle performance, yet little is known about how well these technologies work across social differences. We investigate the biases of deep convolutional neural networks (dCNNs) in scene classification, using nearly one million images from global and US sources, including user-submitted home photographs and Airbnb listings. We applied statistical models to quantify the impact of socioeconomic indicators such as family income, Human Development Index (HDI), and demographic factors from public data sources (CIA and US Census) on dCNN performance. Our analyses revealed significant socioeconomic bias, where pretrained dCNNs demonstrated lower classification accuracy, lower classification confidence, and a higher tendency to assign labels that could be offensive when applied to homes (e.g.,"ruin","slum"), especially in images from homes with lower socioeconomic status (SES). This trend is consistent across two datasets of international images and within the diverse economic and racial landscapes of the United States. This research contributes to understanding biases in computer vision, emphasizing the need for more inclusive and representative training datasets. By mitigating the bias in the computer vision pipelines, we can ensure fairer and more equitable outcomes for applied computer vision, including home valuation and smart home security systems. There is urgency in addressing these biases, which can significantly impact critical decisions in urban development and resource allocation. Our findings also motivate the development of AI systems that better understand and serve diverse communities, moving towards technology that equitably benefits all sectors of society.

Problem

• Investigates socioeconomic biases in deep learning scene recognition.
• Reveals lower accuracy and offensive labeling in low SES images.
• Highlights need for inclusive datasets to ensure equitable AI outcomes.

Innovation

• Used deep convolutional neural networks for scene classification
• Analyzed socioeconomic bias using global and US image datasets
• Proposed inclusive training datasets to mitigate bias