PredPol: Artificial intelligence plus Big Data equals less crime

Computers may not exactly “think” yet but machine learning – where a computer can take an action without being programmed for it – is becoming a formidable tool in the fight against crime.

Crime fighting has entered the tech age. The same crime data now available to the public can produce a wealth of mineable digitized information. Each case has a detailed report that answers basic questions surrounding how and where a crime occurred, in the hope that victimization will not be repeated. With enough historical crime data, environmental trends can be inferred.

What would happen if this mother lode of Big Data were turned into a digital resource to help law enforcement prevent crime?

We already know: the crime rate goes down.

Rod has joined the board of an innovative company called PredPol, a company named for the field that it pioneered: predictive policing. It’s based on the idea that by asking the right questions and analyzing large amounts of crime data, you can more accurately predict where and when crimes may happen. This allows law enforcement agencies to more effectively deploy their resources so they can be in the right place at the right time.

How does it work?

PredPol’s software is a powerful tool that allows law enforcement to shift from reactive to preventive policing. The company has developed an algorithm, based on patterns of previous criminal behavior, that translates three key facts – type of crime, location and time – into customized predictions for individual law enforcement agencies. Significantly, it does not include who committed those crimes. PredPol does not use personally identifiable information about individuals or groups of individuals. That avoids concerns about personal liberty, privacy and profiling, as well as other biases that could emerge if gender, race and other personally identifiable information were considered.

PredPol gives law enforcement the ability to adopt a deterrence-based prevention strategy by providing specific daily predictions – in areas as small as 500 by 500 feet – of where and when crimes are most likely to occur, well enough in advance to plan for them and to deploy resources strategically. Placing law enforcement officers more visibly in high-risk areas, as determined by the algorithm, increases the effectiveness of policing efforts and maximizes the impact of resources such as personnel, situational awareness, knowledge, and experience.

Significant results and benefits

PredPol’s algorithm based predictive analysis is already being credited with yielding dramatic results in preventing crimes as diverse as gun violence, burglary, vehicle theft, drug incidents, gang-related activity, arson and vandalism – even traffic accidents.

Over 60 police departments on three continents are already using PredPol’s technology and, according to a 2012 survey on Forbes.com, 70 percent of almost 200 police agencies planned to implement or increase their use of predictive policing technology in the following two to five years.

It’s easy to see why. When predictive policing is fully in place, the technology can enable law enforcement agencies to reduce targeted crime as much as 20-30 percent.

In PredPol’s first year of use in Santa Cruz, California, where the company is based, burglaries declined by 11 percent and robberies by 27 percent. In one district of Los Angeles, predictive policing reduced crime by 20 percent from January 2013 to January 2014. Similar results are emerging in Atlanta, where predictive policing has already produced a 19 percent drop in overall crime.

The benefits of such significant crime reduction are obvious: less violence, fewer victims, reduced costs, and the enormous social and economic benefits and sense of wellbeing that come with safer and more stable communities.

Down the road

Big Data can be a major strategic asset for an organization. Making sense of so much information and using it proactively can be made simpler and more effective with advanced analytics technology like machine learning.

Rod is involved in developing two other highly innovative companies that rely on the powerful combination of Big Data and algorithms to develop more effective methods to fight cyber crime. One applies machine learning to prevent hackers from using automated scripts to exploit corporate account and financial systems. The other is introducing a new generation of anti-spam protections using a form of artificial intelligence. Initial trials in both of these companies have indicated a high degree of success is likely, suggesting a bright future for machine learning as a sophisticated crime-fighting tool.