Is Your Neighborhood Safe? Place-based Stigma in Large Language Models' Urban Safety Judgments

• URL: http://arxiv.org/abs/2608.26188v1
• Date: Sun, 23 Aug 2026 02:40:41 GMT
• Title: Is Your Neighborhood Safe? Place-based Stigma in Large Language Models' Urban Safety Judgments
• Abstract summary: We probe seven instruct-tuned models that dissociate name from geography.<n>We find ratings are nearly flat under coordinates for six of seven models.<n>Names lower safety ratings more for neighborhoods with higher shares of the locally dominant marginalized group.
• Score: 13.972342642403186
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models are increasingly used to inform safety decisions in cities, such as where it is safe to walk, rent, or travel. We ask whether such judgments track measured risk or the patterns attached to an urban neighborhood's name. We probe seven instruct-tuned models under three conditions that dissociate name from geography: coordinates-only, name-only, and name+coordinates, across 186 neighborhoods in Los Angeles and Chicago, joined to violent crime and American Community Survey data. First, ratings are nearly flat under coordinates for six of seven models, while names carry most between neighborhood variation and are moderately calibrated to violent crime; only at frontier scale does the coordinate channel show appreciable variation. Second, names lower safety ratings more for neighborhoods with higher shares of the locally dominant marginalized group (percent Black in Chicago, percent Hispanic in Los Angeles), and this name effect tracks demographic share in all seven models and both cities. In Los Angeles, where demographic share and crime are more separable, the effect survives controls for crime and income and is confirmed by crime-matched pairs. An enforcement-elasticity analysis further shows that over-caution tracks near-fully-reported homicide rather than discretionary, deployment-driven offenses. Third, the effect scales with geographic knowledge: models that better distinguish real neighborhoods apply more demographic stereotype to them. Because neighborhood names carry both genuine crime signal and demographic stereotype, removing names reduces both bias and accuracy. We discuss implications for deploying LLMs in advice and decision-support settings.

Related papers

• A Digital Shadow for Modeling, Studying and Preventing Urban Crime [0.0]
  Around 80 percent of the world's population lives in countries with high levels of criminality.<n>This paper presents the development and validation of a digital shadow platform for modeling and simulating urban crime.
  arXiv  Detail & Related papers  (2025-01-08T11:31:39Z)
• What's in a Name? Auditing Large Language Models for Race and Gender Bias [45.1187517058961]
  We employ an audit design to investigate biases in state-of-the-art large language models, including GPT-4.<n>We find that the advice systematically disadvantages names that are commonly associated with racial minorities and women.
  arXiv  Detail & Related papers  (2024-02-21T18:25:25Z)
• Large Language Models are Geographically Biased [47.88767211956144]
  We study what Large Language Models (LLMs) know about the world we live in through the lens of geography. We show various problematic geographic biases, which we define as systemic errors in geospatial predictions.
  arXiv  Detail & Related papers  (2024-02-05T02:32:09Z)
• Ranking the locations and predicting future crime occurrence by retrieving news from different Bangla online newspapers [2.3979112009934163]
  We have come up with an approach which can give an approximation to people about the safety of a specific location. Our approach relies on different online Bangla newspapers for crawling the crime data.
  arXiv  Detail & Related papers  (2023-05-18T04:19:26Z)
• Spatial-Temporal Hypergraph Self-Supervised Learning for Crime Prediction [60.508960752148454]
  This work proposes a Spatial-Temporal Hypergraph Self-Supervised Learning framework to tackle the label scarcity issue in crime prediction. We propose the cross-region hypergraph structure learning to encode region-wise crime dependency under the entire urban space. We also design the dual-stage self-supervised learning paradigm, to not only jointly capture local- and global-level spatial-temporal crime patterns, but also supplement the sparse crime representation by augmenting region self-discrimination.
  arXiv  Detail & Related papers  (2022-04-18T23:46:01Z)
• Measuring and Modeling Neighborhoods [0.0]
  We develop an open-source survey instrument that allows respondents to draw their neighborhoods on a map. We propose a statistical model to analyze how the characteristics of respondents and local areas determine subjective neighborhoods.
  arXiv  Detail & Related papers  (2021-10-26T20:43:41Z)
• The effect of differential victim crime reporting on predictive policing systems [84.86615754515252]
  We show how differential victim crime reporting rates can lead to outcome disparities in common crime hot spot prediction models. Our results suggest that differential crime reporting rates can lead to a displacement of predicted hotspots from high crime but low reporting areas to high or medium crime and high reporting areas.
  arXiv  Detail & Related papers  (2021-01-30T01:57:22Z)
• More crime in cities? On the scaling laws of crime and the inadequacy of per capita rankings -- a cross-country study [0.0]
  Crime rates per capita are used virtually everywhere to rank and compare cities. We demonstrate that using per capita rates to rank cities can produce substantially different rankings from rankings adjusted for population size.
  arXiv  Detail & Related papers  (2020-12-30T23:36:28Z)
• Discovering Underground Maps from Fashion [80.02941583103612]
  We propose a method to automatically create underground neighborhood maps of cities by analyzing how people dress. Using publicly available images from across a city, our method finds neighborhoods with a similar fashion sense and segments the map without supervision.
  arXiv  Detail & Related papers  (2020-12-04T23:40:59Z)
• Socio-economic, built environment, and mobility conditions associated with crime: A study of multiple cities [9.78342936850961]
  We propose a Bayesian model to explore how crime is related to socio-economic factors. We find that the combined use of socio-economic conditions, mobility information and physical characteristics of the neighbourhood effectively explain the emergence of crime. We show that the socio-ecological factors of neighbourhoods relate to crime very differently from one city to another.
  arXiv  Detail & Related papers  (2020-04-13T08:36:59Z)
• Measuring Spatial Subdivisions in Urban Mobility with Mobile Phone Data [58.720142291102135]
  By 2050 two thirds of the world population will reside in urban areas. This growth is faster and more complex than the ability of cities to measure and plan for their sustainability. To understand what makes a city inclusive for all, we define a methodology to identify and characterize spatial subdivisions.
  arXiv  Detail & Related papers  (2020-02-20T14:37:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.