Technology
Predictive Policing and the Erosion of Civil Liberties
Quick fact
Several U.S. cities, including Los Angeles and Chicago, have used predictive policing tools, but some have discontinued them due to complaints about racial bias and ineffectiveness.
Why this is interesting
Imagine if a computer could predict where the next crime will happen—should the police go there? And what if the prediction is wrong?
Read the full explanation
Understanding Predictive Policing and the Erosion of Civil Liberties
Predictive policing works like a weather forecast for crime. Software analysts look at past crime reports, times, locations, and even social media posts to estimate where and when future crimes are likely. The police then send patrols to those 'hot spots.' The idea is to deter crime before it happens. But here's the catch: the system uses historical data that is already biased by past policing practices. If police historically focused on certain neighborhoods, the data will show more crime there, and the algorithm will send police back there again. This creates a loop where the same communities are continuously targeted, even if crime rates are similar elsewhere.
A deeper explanation
The mechanism behind the erosion of civil liberties lies in the feedback loop. Algorithms are trained on historical records, which reflect not actual crime rates but the outcomes of previous police deployments, citizen reporting habits, and systemic biases. When the algorithm predicts 'more crime' in a neighborhood, police patrol there more, leading to more arrests and reports, which feeds back into the next round of predictions. This cycle amplifies existing disparities. Furthermore, predictive policing often uses opaque algorithms that are not open to public scrutiny, making it difficult for citizens to challenge decisions. It also relies on personal data—such as location tracking and social media activity—raising privacy concerns. The presumption of innocence is threatened when individuals are flagged as 'likely to commit a crime' based on statistics, rather than on evidence of wrongdoing. Overall, the technology, intended to be neutral, can systematically undermine equal protection under the law.