In the prisons of North Rhine-Westphalia, machine learning should help prevent suicide attempts by the inmates. To this end, the country launched a nationwide unique research project, announced NRW Justice Minister Peter Biesenbach (CDU) on Tuesday in Dusseldorf. The aim is to develop a computer program that recognizes suicidal behavior patterns at an early stage and can be used accurately in the video surveillance of vulnerable prisoners.
As an example, the ministry cited conspicuous behaviors such as movement patterns in a strangulation attempt or the use of dangerous objects such as knives.
The contract for the development of the software was obtained by a company from Chemnitz. There, a test room will be built that meets the requirements of a typical prison room. The knowledge gained there would be backed up by additional tests in the prison (JVA) Dusseldorf in real operation, said Biesenbach.
During the pilot phase, vulnerable prisoners will be monitored by video cameras next year. Using artificial intelligence, the technique should detect suicidal actions and sound the alarm when it detects them. Since the eyes of the employees could not be everywhere, the judiciary in North Rhine-Westphalia set on the technical progress, explained the Minister. The artificial intelligence should not replace the previous preventive measures, but complement.
A so-called "initial screening" is already prescribed in the NRW institutes in order to better identify suicidal detainees. Those considered to be at high risk are continuously monitored in a specially secured cell. Less acute cases are controlled at intervals. In the past three years, the number of suicides in the 36 detention centers in NRW has been declining. According to the NRW Ministry of Justice, eleven prisoners died in 2018; It was 13 in 2017 and 19 in 2016.
