The article discusses the use of Geolitica, a crime prediction software, highlighting its racial and socio-economic bias in directing police patrols. An analysis across 38 jurisdictions revealed that the software unfairly targeted neighborhoods with higher concentrations of low-income, Black, and Latino residents. Despite these predictions, arrests in these areas remained constant, indicating a lack of correlation between predicted crimes and actual law enforcement outcomes. The investigation faced limitations due to insufficient data on police response to predictions, making it difficult to assess the software's actual effectiveness or accuracy.
Our analysis revealed that the Geolitica software, used for predicting crime, disproportionately directs law enforcement to neighborhoods with high low-income, Black, and Latino populations.
Despite predictions made by Geolitica, the rates of arrest in predicted areas did not show a significant correlation with the software's crime predictions.
Collection
[
|
...
]