Scientists have developed an algorithm that predicts the commission of crimes

Scientists from the University of Chicago in the United States have developed algorithmEvent-level prediction of urban crime reveals a signature of enforcement bias in US cities, which allows you to predict the commission of crimes in a particular area. It works on the basis of artificial intelligence, which studies the spatio-temporal relationships from reports of incidents - murders, attacks, robberies, car thefts, and so on.

All this basis is data from open sources. The algorithm tracks patterns in them in time and place, and as a result, it can predict the commission of a crime within the next week with an accuracy of about 90%.

The AI ​​model divides the city into equal squares with a side of about 300 meters. And the data with the probability of some kind of incident is determined for each of them.

The algorithm has already been successfully tested in Chicago, Atlanta, Detroit, Los Angeles, Portland, San Francisco and other cities.

The creators of AI themselves believe that it is not worth using such systems directly to ensure law and order. After all, an increase in the number of police officers in areas of the city where crime is expected will lead to a change in the conditions of modeling. And this can already reduce the efficiency and accuracy of prediction, violating all conditions.

At this stage, such predictive models make it possible not so much to prepare for a crime as to identify the causes changes in their number, as well as to improve control over the work of the police themselves, identifying systemic violations in their work and bias.