Richard Berk, a professor of criminology at the University of Pennsylvania, has developed an algorithm and software that he claims can accurately predict which criminal offenders on parole or probation are more likely to commit a serious crime, including murder.
In the past, it was primarily a parole officer’s judgment, along with a criminal’s past record, used to influence the decision regarding the level of supervision necessary for each person. Admittedly, human judgment is often less than perfect, however, equally, individual intuition or ‘gut feel’ about another human being can in many cases, not be reproduced or replaced by computer software technology.
What can certainly help in predicting criminal behavior, is technology that can augment the personal or subjective side of evaluation, using the raw processing speed and accuracy of computer software. Berk’s software technology incorporates a database containing more than 60,000 crimes, including homicides. The software program also includes approximately two dozen variables, utilizing not only empirical data such as age and criminal record, but also geographic location, type of crime, and the relative age of each person when they committed the crime.
It’s very important to distinguish, that this new technology does not predict actual crimes, but rather, only the tendency or likelihood of which criminals might commit crimes again. Consequently, this ability will directly translate into the level of supervision each person on parole or probation gets assigned by the courts.
New software enhances police profiling efforts
Criminal profiling is nothing new in law enforcement, however, using technology as a foundation to assist in the evaluation of people and probability estimates, is much newer and has proven to be more accurate in predicting behavior patterns of criminals. As with the use of any technology, there is always the acknowledgement that output is less than 100% perfect. Regarding Berk’s software technology and use, the increased accuracy and additional information directly translates into an increased knowledge base, which in turn, results in enhanced decision-making.
As a result of the high level of success this technology has yielded to-date, statistical crime prevention efforts are likewise scheduled to be used across the nation, specifically in Washington, DC, Boston, New York City, Los Angeles and Chicago.
Sometimes, the good guys do finish first.
Comment and add to the story without registration, but keep the comments meaningful please. Links are not accepted.
