Sophisticated crime software helps police predict violent offences

Published 27 July 2010

Minority Report”-style technology being tested by two British forces following success in the United States; the system, known as CRUSH (Criminal Reduction Utilizing Statistical History) evaluates patterns of past and present incidents, then combines the information with a range of data including crime reports, intelligence briefings, offender behavior profiles, and even weather forecasts

Two British police forces have begun trials of a sophisticated computer software package which aims to boost their efficiency by predicting where and when future crimes will take place. The system, known as CRUSH (Criminal Reduction Utilizing Statistical History) evaluates patterns of past and present incidents, then combines the information with a range of data including crime reports, intelligence briefings, offender behavior profiles, and even weather forecasts. This is used to identify potential hot spots and flashpoints, so police forces can allocate resources to areas where particular crimes are most likely to occur.

The Guardian’s Tony Thompson writes that the technology, which belongs to a fast-growing field known as “predictive analytics,” is being tested secretly in the United Kingdom following a successful long-term trial in Memphis, Tennessee, where the police department credits CRUSH as the key factor behind a 31 percent reduction in overall crime and a 15 percent fall in violent crime. The system has also been credited with improving morale among officers of the Memphis police by boosting arrest rates and helping them to feel as if they are “making a difference.”

This is more of a proactive tool than reacting after crimes have occurred. This pretty much puts officers in the area at the time that the crimes are being committed,” said John Williams of the Memphis Crime Analysis Unit.

The software behind the system has been developed by global computing giant IBM, which sees the increasing use of analytics as a massive growth area for the future and has invested more than $11bn in the field in the past four years. The names of the two U.K. forces using the software have not been revealed.

Thompson writes that the increasing use of predictive analytics by law enforcement and judicial agencies around the world has sparked inevitable comparisons with the Tom Cruise science fiction film “Minority Report” in which police “pre-crime” units use predictions made by psychics to apprehend potential offenders before their crimes have taken place.

Earlier this year the Ministry of Justice began using predictive analytics to assess the data held within its Offender Assessment System and help predict which prisoners due for release were most likely to re-offend based on circumstances such as accommodation, education, relationships, financial management and income, lifestyle and associates, drug and alcohol misuse, emotional well-being, behavior and attitudes.

In Florida, the U.S. Department of Justice recently began using the same software to help predict which young delinquents were likely to go on to become repeat offenders, placing those flagged up by the computer system into specific prevention and education programs aimed at ensuring they remained on the straight and narrow.

Critics say the use of such technology is an affront to human rights and could destroy centuries of legal precedent, leading to a generation who are innocent only until predicted guilty. While supporters point out that at present, such analysis and decisions are made by individuals prone to making mistakes and unable to take into consideration the wealth of information a computer can deal with.

According to Mark Cleverley, head of government strategy at IBM, CRUSH simply enhances and improves the efficiency of existing practices. “What the technology does is what police officers have always done, sometimes purely on instinct — looking for patterns to work out what is likely to happen next. What is different is the scale on which the system operates and the speed at which the analysis takes place.”

Last week Julie Spence, chief constable of Cambridgeshire police, complained that increasing paperwork and red tape were reducing the amount of time her officers could spend out on patrol. IBM hopes that police forces around the world, struggling with limited resources, will be eager to adopt a system which allows them to maximize the efficiency of officers on the beat.

According to Cleverley, the company is now refining the system to enable it to sample data from an even wider range of sources and process the results faster. “At some point in the future we hope to include analysis of feeds from CCTV cameras and public sources from the internet such as Facebook posts.” Had such a system been in place it might have prevented Raoul Moat’s rampage. Prior to his shootings he had issued threats on his Facebook account.