Development of and Concerns Regarding Predictive Policing Practices
Predictive policing is the application of analytical software to identify likely targets for police intervention and prevent crime. The last 8–10 years has brought the application of analytical tools and artificial intelligence to enable such predictions based on large sets of data. Proponents of predictive policing say that it better processes information without human bias, preventing police officers’ acting out of prejudice, or even distraction, in order to allocate police resources more efficiently and equitably. However, there is no widespread evidence that the initiatives improve community safety, and numerous advocacy groups and legal challenges have called attention to the dangers of predictive policing in terms of reproduction of biases, civil rights violations, and lack of transparency.
Key findings
- Predictive policing programs utilize algorithms to allocate law enforcement resources to areas and persons identified as having a higher risk for crime, despite a lack of evidence showing the efficacy of these determinations.
- Commercially-produced and in-house predictive policing software varies greatly in design and is not subject to standardization or government accountability.
- Predictive policing programs rely on biased historical data, threaten the civil rights of community members, and do not provide for transparency or accountability regarding their practices.
- For these reasons, predictive policing tools should not be used in any law enforcement agency. If agencies are using these tools, they must be subject to frequent evaluation and practice transparency regarding data.
