Predictive policing, a method that uses data analysis to forecast criminal activity, is gaining traction.

In Oregon, understanding how these systems work and the legal implications is crucial. We’re keeping a close eye on these developments and will continue to report on them through the Corbridge Law blog.

How Predictive Policing Works 

Predictive policing models aim to determine where and when to deploy law enforcement resources more efficiently. They rely on several key data inputs:

● Crime Reports: Details like offense type, time, location, and “modus operandi.”

● Calls for Service: Frequency and clustering of emergency and non-emergency calls. 

● Arrest and Citation Records: Location and type of arrests, with a critical eye towards preventing profiling. 

● Environmental Data: Factors such as land use, lighting, and vacant properties. 

● Temporal Variables: Time of day, day of week, and seasonal crime fluctuations. 

● Geographic and Demographic Patterns: Socioeconomic factors and population density. (profiling?) 

These data points are fed into analytical systems that typically fall into two categories: 

● Place-Based Prediction: This identifies “hot spots” they decide are “likely to experience future crime.” 

● Person-based Prediction: A more controversial approach that attempts to identify individuals at higher risk of offending or victimization, which is highly scrutinized for potential biases. (Systems using previously biased data would carry that bias into future prediction, creating a self-fulfilling prophecy.)

The outputs of these systems often manifest as risk or heat maps, guiding the deployment of officers, vehicles, drones, or surveillance systems to enhance efficiency and responsiveness.

Legal and Ethical Considerations in Oregon 

While the promise of increased efficiency is appealing, there are significant legal and ethical challenges, particularly within Oregon’s legal framework: 

●Data Bias: A major concern is that historical crime data can reflect and perpetuate existing biases, leading to over-policing in marginalized communities. 

●Transparency: For public accountability, the algorithms used in predictive policing must be explainable and their outputs subject to public scrutiny. 

●Human Oversight: Oregon law mandates human review, meaning AI systems cannot autonomously direct law enforcement actions. Human judgment remains paramount. 

●Privacy and Data Retention: Any use of surveillance or private data must strictly adhere to Oregon statutes and privacy standards, including ORS provisions. As an example, the AI pilot in Bend, Oregon, currently focuses on generative uses like report drafting.

If expanded to predictive policing, it would likely consider factors such as crime density, calls for service trends, nightlife activity, local event calendars, and seasonal tourism variations.

The Path Forward 

Predictive policing presents a double-edged sword where potential gains in efficiency are provided with significant risks to civil liberties. Oregon agencies must implement robust governance frameworks, conduct thorough bias audits, and ensure transparent public oversight before deploying these predictive systems on a large scale, but is that enough?

At Corbridge Law, we believe in upholding justice and protecting individual rights. We will continue to monitor the implementation and legal challenges of predictive policing to ensure these technologies serve the public good without infringing on fundamental freedoms.