Skip to main content

Finding a Fair and Equitable Use of Citizen Data: The Case of Predictive Policing


Releasing data to the public can help to unleash innovation and generate huge benefits. In the public sector using citizen’s personal information allows the government to deliver services proactively and with greater efficiency. However, the use of citizen data raises several points of concern for whether current safeguards are adequate for preventing misuse. One area where the use of open data deeply affects society is in the criminal justice system.

A hallmark of big data is personalization and segmentation tools that parse out demographic information such as gender, ethnicity, class, age, personality, habits, and geographical attributes to group and profile users. These profiles are then used to inform actions and decisions. One way in which citizens’ data are being used to inform decisions locally is through predictive policing.

Predictive Policing

Predictive policing is the use of analytics through finding associations and patterns in criminal activity. The program weighs several risk factors (i.e. prior arrests, acquaintances, acquaintances background, parole status, etc.). The analysis then reveals names of individuals that are 5, 50, or 500 times more likely than average to be involved in a crime. In cities like Chicago where the National Institute of Justice has funded their Two Degrees of Association program, once a person has been identified, the police will visit them and issue a warning not to commit a crime.

Evidence-Based Sentencing

For individuals who have already committed crimes, in some states, their data are being used in their sentencing. Evidenced-based sentencing is a practice where a judge uses information about the offender’s risk, needs, and responsiveness to rehabilitation to inform sentencing. The information empirically quantifies the risk of the offender’s future threat he or she is likely to impose on the community and the predicted effectiveness of recidivism treatment.

In 2014, U.S. Attorney General Eric Holder wrote in a report to the U.S. Sentencing Commission condemning the use of evidence-based sentencing. In it he acknowledged and praised the value of data analysis and what it offers the criminal justice system. He also highlighted the importance of the human component involved in sentencing. He argued that using offender characteristics such as education level, employment history, family circumstance, and other demographics instead of basing sentencing on the crime committed is a ‘dangerous concept’ and far reaching. In implementing such practices, two individuals could commit the same crime but receive very different punishments.


He also noted that such practices ultimately raise questions of constitutionality, due to the use of group-based characteristics and suspect classifications. In Bearden v. Georgia in 1983, the Georgia Court of Appeals rejected the state’s intentions of revoking the plaintiff’s probation because of his inability to find employment. The rationale was that unemployment heightened the plaintiff’s risk for reoffending. The court found that classifying anyone as dangerous because they are poor is, in effect, punishing poverty and not constitutional.  Attorney General Holder concluded that the trend of using predictive analytics to improve programming and efficacy in other parts of criminal justice such as evidence-based diversion programs will grow but warns against taking them into arenas they don’t belong such as sentencing.

Whether or not these programs work, there is no doubt that they impact society. Unbeknownst to many citizens, their personal information is being used against them. Utilizing analytics will have an unequal impact on more vulnerable populations. The public sector has yet to find an equitable and fair balance to these issues but challenges are on the horizon. Public agencies such as the National Institute of Justice have been in the forefront of opening dialogue on the issue through its annual Predictive Policing Symposium. However, it will be up to policymakers and criminal justice leaders to find an equitable solution to a pressing problem without creating more problems for the future. 



Kendra L. Smith

Kendra L. Smith is a doctoral candidate in the School of Community Resources and Development within the College of Public Programs at Arizona State University.


Get daily updates from Brookings