Technology

Artificial intelligence has learned to look for robbers by pattern of behavior.

Finding similarities between crimes,committed in different areas of large cities, the police can quickly find and arrest criminals. According to Digital Trends, to recognize the “handwriting” of robbers, investigators had to manually compare many protocols, which took a lot of time and effort. It turned out that for this task, the New York Police Department has been using artificial intelligence for three years, which can cope with it much faster.

The software is called Patternizr, andIts creators are Assistant Commissioner of the New York Police Department Evan Levine and his colleague Alex Chochlas-Wood. It took about two years to develop, and the data collected for one decade was used for training. To identify common features of robberies, the algorithm pays attention to the door breaking method, the list of stolen items and the scene of the incident.

The algorithm began to be applied in 2016, but aboutit became known only now from an article in the Informs Journal, where developers discussed their work. Its effectiveness has been proven many times - for example, he was able to link together two robberies that were committed in two different parts of New York with a difference of two weeks. The criminal issued his unusual weapon - a syringe, which he intimidated the staff of the Home Depot store.

A manual matching method would take muchmore time and not the fact that the officers could find a match. In the worst case, the robber would still remain at large. Thanks to Patternizr, instead of many hours of work, the investigators needed to make just one mouse click.

It is noteworthy that artificial intelligence is alreadyhelped the police in a more difficult and responsible case - making a decision when making a sentence. About what can be dangerous artificial intelligence in a court case, you can read in our special material.

What can you say about the new method of finding burglars? Write your opinion in the comments under the news, or in our Telegram-chat.