NYPD New Software, AI Catches Criminal With Pattern Recognition


The New York Police Department (NYPD) is known to be the first department in the history of the United States to employ a pattern-recognition software to identify patterns or similarities between crimes occurring over precincts.

The Associated Press reported that the machine-learning AI was built in-house over a period of two years and the department has been using it since 2016 but has only brought it in front of the public eye recently.

According to Chohlas-Wood, the real advantage of the software is to reduce the amount of legwork and to employ the expertise of detectives and their experience in going through a much smaller list of possibilities.

10 Years Worth of Data

The software has been fed with ten years worth of data, manually identified and solved. It utilizes the data linked to each crime like the distance between the crime locations, type of goods targeted, the method used to enter the premises of the crime and much more like this.

Analyzing past events helps the software in the easy identification of patterns in the more recent criminal activities. Before the software, investigators mainly focused on the activities concerned with their precinct. Communicating over precincts to discuss patterns and similar behavior was a hassle.

With the introduction of this software, that has been made easier and more accessible.

In Action

According to the Association Press report, the software has been successful in linking two crimes that occurred several weeks apart and included a man trying to rob two Home Depot stores by threatening the staff and workers with a syringe. What might have taken hours to observe and analyze by the human brain and mind was done in a couple of minutes by the program? Which is what makes it more efficient in employing it in further private and public investigations.

If not for the software then maybe it would not have been possible for the NYPD to arrest the man thus enabling a criminal to roam in the streets.

The old formula, according to Evan Levine and Chohlas-Wood, the developers of the software, was so time-consuming that it took hours and days to sift through the mountain of reports manually looking for one link.

The software not only helps in identifying patterns but also helps in finding more clues, thus saving the investigators precious time. Furthermore, it helps a cop in tracking down a person who has been operating over the seventy-seven precincts.

Leave a Reply

Your email address will not be published. Required fields are marked *