JammedUp News


TECH BIAS: Is software used by cops to predict crime racist against blacks?

May 24, 2016  |  Posted by: JammedUp Staff
TECH BIAS: Is software used by cops to predict crime racist against blacks? GETTY IMAGES

Are you in a legal jam? Find a Lawyer, Bail Bondsman or Private Investigator on JammedUp.

Criminals throughout the United States are now being subjected to a computerized assessment that evaluates the likelihood of reoffense and recidivism.

The point of the evaluations is to generate data on suspects, in which they are assigned scores based on their level of risk to the public. Judges, in turn, use this information to help them during sentencing.

However, the Daily Mail has reported that more recent investigations claim that racial characteristics may play a central role when assigning risk scores.

An investigation, conducted by ProPublica, discovered that the software that police are using to generate ratings may be biased against African-Americans, who have higher rates of recidivism compared to any other group and they are also the largest imprisoned population. The software is believed to be more likely to flag African Americans as higher risk for reoffense while wrongfully identifying Whites as lower risk.

Many are arguing that these risk assessments have become a form of racial profiling and that they do not take into consideration access to services and probation conditions.

The assessments of approximately 7,000 people in Florida between 2013 and 2014 were collected to analyze the likelihood of recidivism over two years.

The report released by ProPublica noted that among those predicted to re-offend over the two year period, only one out of five individuals did so.

Scores for African-American individuals were distributed from 1-10 (with one being considered lowest risk) while Whites’ scores were more likely to be more heavily concentrated on the lower risk part of the spectrum.

One example cited in the report depicted two individuals, one White, and one African American, who were both arrested for drug possession. When taking into account previous offenses, the White suspect was charged with attempted burglary, while the other was arrested for resisting arrest. However, the Black suspect was considered to be at greater risk of reoffense.

When assessed two years later, the Black individual stayed out of trouble, while the “low risk” White offender was arrested a number of times for drug possession.

Correctional Offender Management Profiling for Alternative Sanctions, which is the software used to generate the risk assessments, has disputed the report released by ProPublica, noting that they used inadequate and erroneous statistical methods to conduct the analysis.

A spokesperson told the DailyMail that: “that the conclusions drawn from the results are misleading.”

Get the latest news from the world of crime