Will an Algorithm Help Colombia Predict Crime?

SHARETweet about this on TwitterShare on FacebookShare on LinkedIn

An initiative to create an artificial intelligence system that analyzes massive amounts of data to recognize criminal patterns underscores how Colombia seeks to use advanced technology to respond to crime threats.

The system, which is expected to first be deployed in Bogotá, will use crime data to pinpoint hotspots, identify patterns, and predict where crimes are likely to occur.

To develop the system, authorities in Bogotá teamed with the mathematics faculty at the Universidad Nacional and a private company that specializes in predictive modeling. The project is expected to be finished in less than three years, Semana reported.

Through data from the police, prosecutors, security cameras, emergency dispatchers, surveys, and other sources, the system’s aim will be to detect the “largest patterns to know the when, how and why of criminal acts in Bogotá, and realize interventions at the local level,” said Francisco Gomez Jaramillo, one of the project’s directors and a mathematics professor at Universidad Nacional.

SEE ALSO: Colombia News and Profile

This is just the latest effort by Colombian authorities to incorporate data into their crime-fighting strategies.

Others include the online crime reporting portal Adenunciar, and the use of data crunching software by the Attorney General’s Office to detect false claims made within the country’s public health system. The latter found some three million false claims costing some $724 million Colombian pesos (around $228 million), according to a report by the Attorney General’s Office.

Attorney General Néstor Humberto Martínez has also credited his office’s use of computer-assisted data analysis to detect an armed robbery ring. The group convinced victims that a fictitious military officer had discovered a “caleta,” or stash, of buried drug money, and then offered to exchange the US cash at half the price. When the victims arrived with the money, the group robbed them at gunpoint. Through data analysis, prosecutors identified 58 cases where the same method was used across eight departments of Colombia, Martínez said in a news release.

Colombia has long used DNA in forensic investigations, but the Attorney General’s Office recently asked for funding to create a DNA database for the purpose of crime solving. Martinez stated that such a system would improve the country’s investigative standards by 80 percent. However, the ambitious proposal would have to undergo a long legislative process to receive approval.

InSight Crime Analysis

The use of advanced technology to collect and analyze crime data has shown some success in Colombia. But the proposed artificial intelligence system — while innovative and a potential gamechanger in attacking crime — faces vast difficulties in setting up and comes with some potential drawbacks.

SEE ALSO: Coverage of Security Policy

When describing the system, Gomez Jaramillo said deciding data inputs and designing a reliable algorithm is an enormous hurdle.

“If the [data] are not relevant,” he told Semana, “then we would have nothing to interpret.”

There is also the concern of machine bias. US authorities have used algorithms to aid judges in determining sentences for crimes by providing scores on the likelihood that the person will commit another crime. Yet an investigation by news organization Pro Publica found that such algorithms were “remarkably unreliable in forecasting violent crime,” and showed bias against black people.

Countries in Latin America have deployed ineffective artificial intelligence systems. Uruguay, for example, implemented a predictive policing software, but the program was scrapped after three years due to the system’s “high degree of opacity and its

potential to reinforce discrimination and exclusion,” according to a World Wide Web Foundation report.

Even with such issues, it is clear that pattern-recognition tools that can analyze massive data sets are the future of policing, and Colombia appears to be on the cutting edge.

SHARETweet about this on TwitterShare on FacebookShare on LinkedIn