National / Crime & Legal

Kanagawa police to launch AI-based predictive policing system before Olympics

Kyodo

The Kanagawa Prefectural Police plan to become the first in the nation to introduce predictive policing, a method of anticipating crimes and accidents using artificial intelligence, sources said Sunday.

The Kanagawa police will seek research expenses under the prefecture’s budget for fiscal 2018 starting April, hoping to put a predictive policing system in place on a trial basis before the 2020 Tokyo Olympics, prefectural government sources said.

A system that can determine whether a single perpetrator is behind several crimes, predict an offender’s next move and detect where and when crimes or accidents are likely to occur would help police officers investigate crimes and prevent some from happening, they said.

It would allow them to patrol the suggested places at the most likely times to ensure safety and would also help speed up probes, the sources said.

The AI-based system would employ a “deep learning” algorithm that allows the computer to teach itself by analyzing big data. It would encompass the fields of criminology, mathematics, and statistics while gathering data on times, places, weather and geographical conditions as well as other aspects of crimes and accidents.

It may also tap information gleaned from social media.

The Kanagawa police began studying the feasibility of using such a system last year and plans to begin joint research with private-sector entities this spring before putting a system into practice, the sources said.

They have already used a system to list areas with frequent crime, but it has fallen short of helping to make extensive predictions.

The National Police Agency set up an advisory panel in December to discuss how it should make use of AI.

Predictive policing is already in use in the United States, where critics are raising concerns that it can be used to violate human rights.

Many police forces in the U.S. use the predictive policing software PredPol, which its website says originated from research by the Los Angeles Police Department and the University of California, Los Angeles.

PredPol predicts areas where a crime is likely to occur by analyzing data about past crimes.

The American Civil Liberties Union and 16 other organizations issued a joint statement in August 2016 condemning the police for making computerized determinations the reason for questioning people or making arrests.

The statement criticized the system for allegedly promoting prejudice against certain communities and residents.

Toyoaki Nishida, professor of information science at Kyoto University graduate school, said preventive measures would be possible if a hypothesis that crimes are concentrated at particular times and places proved correct.

But using such a method may have negative aspects, such as frequent police patrols of the same areas, he said. Such a system would first need to be accepted by residents before being put into use, he added.