Can a Computer Predict Crime?


Can machines predict where crimes occur before human beings do? Will the technology be used in unbiased ways? How will these devices affect privacy?

As concerns grow over what are called predictive-policing methods and their effect on Americans, one thing is clear: these new crime-fighting tools are not only being implemented around the country; they are increasing in sophistication and scope.

Technology such as PredPol, which runs crime reports through an algorithm to discover locations with a high probability of crime, have already been adopted by police departments in Miami, Seattle, Los Angeles, and Santa Cruz.

The Chicago Police Department in 2013 created a “heat list” which logged 400 people through an algorithm thought to expose future criminals. Its effectiveness and accuracy remain unclear.

In New Orleans, a company called Palantir applied technology to identify approximately one percent of the city population who will be likely to inflict or be victims of gun violence.

The use of such crime-fighting tools has sparked debate among academics, legal experts and privacy advocates.

Law professor Andrew Ferguson of the University of the District of Columbia told a symposium at the Association of American Law Schools annual convention in New York on Friday (January 8) that predictive policing devices can “distort the law” regarding the Fourth Amendment prohibition of illegal search and seizures.

“This [technology] is less about predictions and more about risk,” he said. “It’s more about identifying vulnerabilities in a community as opposed to being able to actually predict a particular crime.”

Andrea Roth, a law professor at the University of California-Berkeley, told the “Predicting Trouble: Risk, Technology, and a Data Driven Criminal Justice System” forum that police departments now have “rouge” DNA databases at their disposal.

These forensic labs are not controlled by any laws and don't meet the FBI's CODIS (Combined DNA Index System) standard. The information in these databanks contain genetic profiles of criminal suspects often obtained from low-level defendants who offer their DNA as part of a plea deal or in exchange for having charges against them dropped, she said.

In California's Orange County, Calif., Roth said, the District Attorney's office has more than 120,000 DNA samples in the databank, many genetic traces given as a “trade-off” to avoid imprisonment, she said. Roth, who opposes such databases, believes these low-level offenders are swapping being imprisoned for the risk of being surveilled for future crimes.

“Public prosecutors are trading in this power to vindicate certain norms in exchange for the power of genetic surveillance over the defender (who is) at risk for future, more serious crimes,” she continued. “In essence, the DA’s office is focusing on identifying the more serious future offenders rather than on punishing that serious crime.”

But some predictive methods and studies have proven to be accurate and efficient.

Shima Braughman, a law professor at the University of Utah, studied whether perpetrators would be detained or not during pre-trial hearing periods and found the cases easy to forecast.

“It was really possible to predict,” she said.

Perpetrators of robbery, burglary and motor vehicle theft, for instance, were most likely to be re-arrested in a pre-trial period, she said. Individuals arrested for drug possession and drug trafficking were least likely to commit violent crimes upon release. And those charged previously with three or more violent crimes were statistically likely to commit another violent crime.

Braughman believes predictive models are not altogether new.

“Judges predict all the time,” she said.

During pre-trial bail hearings, for example, a judge decides whether or not to release an individual based on a calculation of that individual's future actions.

“They’re predicting if this person going to commit a crime or not,” Braughman continued. “Whether (prediction is) good or bad, judges are doing it regardless. “

The only difference with the new analytical models is that they give judges the option to base their predictions on hard data, rather than the instinct that comes from long experience on the bench, she said.

According to Christopher Slobogin, a law professor at Vanderbilt University, there’s been a meteoric rise in evidence-based policing, or policing that uses empirical data to identify and address crime risks. In 2012, risk-assessment instruments were used nationwide to analyze nearly 80,000 parole applications, Slobogin said.

Two examples of such instruments are the Violence Risk Appraisal Guide (VRAG) and HCR-20, both of which utilize professional guidelines for assessing and managing violent individuals who enter the criminal justice system. Scoring indicators measure an individual based on factors ranging from mental health and past criminal history to marital status, in order to determine whether an individual is a risk to himself or others if he or she is not confined.

Although VRAG was considered a state-of-the-art tool when it was first tested with white Canadians two decades earlier, some consider it less valuable for black men in the contemporary United States who make up the majority of country's prison population. There was also a 20 to 30 percent incidence of false positives in some cases throughout tests taken in America, he said.

“They [administrators of the tool] also make very clear that many risk factors they rely upon have nothing to do with the perpetrator,” he said.

Yet despite its flaws, Slobogin believes predictive policing offers a useful model.

“I think that the move toward evidence-based sentencing is good, and we ought to be doing this,” he said. “I’m not sure the alternatives to this are any better.”

Other experts in the field have also pushed to increase the number of experiments in order to help change policy and policing.

David Abrams, a law professor at the University of Pennsylvania, said critics who oppose testing redictive policing in law enforcement are short-sighted.

“I can’t really grasp the counter-arguments to running experiments so we can learn about the world when we don’t know the answer —and we don’t know the answer,” he said.

Henrick Karoliszyn is an award-winning journalist based in New York City, specializing in criminal justice reporting. A former staff writer at the New York Daily News and New Orleans Times-Picayune, his work has appeared in The New York Times, the Wall Street Journal, Aeon Magazine, and Fusion. Henrick welcomes comments from readers.

Comments are closed.