Using AI to Detect Future Crimes
We all know that crimes aren’t something new these days. But what’s really interesting is how some countries use artificial intelligence to prevent crimes before they happen. And this trend has spread all over the world.
Artificial intelligence is a broad-spectrum tool. Things like driving autonomous cars, facilitating political decision making or helping to improve our health are some of its utilities. And all sectors of crime will be pursued by this technology. Now the images of the film Minority Report no longer seem utopian to us.
Artificial Intelligence to combat crime
It is already used to manage police resources, to find false reports, to solve crimes and even to guide criminals to ensure that they do not return to commit their own. We took a tour of several of the IA instruments used for crime prevention.
In 2018, the Netherlands published a program to help resolve cases that were long ago inconclusive. But they hope that with the help of the IA these cases will be reinvestigated. At least some of them, since in many of them political data is scarce. And in many cases, most of that data is still handwritten.
His name is Sweetie and he simulates a girl chatting with criminals. In just 10 weeks, Sweetie was able to find 20,000 pedophiles and identify at least 1,000. The Australian, Scott Robert Hanses, was the first to be arrested. This event took place in 2013. The project is being used by police officers around the world.
With the help of European Union funds, VALCRI (Visual Analytics for Sense-making in CRiminal Intelligence analysis) began to be developed in 2017. It is a program that through AI and visual analysis in conjunction with closed circuit surveillance is able to identify suspects. It is still developing but promises great progress and results.
The AI watches our streets and protects us.
The use of cameras with artificial intelligence is very frequent and has been working on it for years. Google image search is a good example. Another is the photographic retouching that we were able to obtain on mobile devices such as the Huawei Mate20 Pro or Huawei P20 Pro. Able to select advanced settings when shooting.
In recent years a police team has achieved the breakthrough, with many achievements, of Veripol. A system that is capable of distinguishing true from false reports by the way they were written. A false statement not only involves the wasted time of the officer who takes it. But also the whole system that is activated to investigate the event. Just initiating the procedure of investigating a case is already a great investment in time, effort and money. But if we also continue with them and they turn out to be a fiasco, those resources are exponential.
In order to carry out the first step, 500 allegations were taken that were true and some that were false. An IA dealt with taking common features and differences between the two types of complaints. And it’s these last ones he pays the most attention to.
When there is a report of robbery with violence and intimidation, or theft with tug of war, it is written in the police station. Then it is compared with the database to decipher, in a line from 0% to 100%, as certain estimates the artificial intelligence. Of course this evidence is not conclusive and the complaint is supported by evidence in situ or witnesses. But it’s certainly a great help tool.