A Canadian artificial intelligence company develops a technology capable of predicting suicides by studying behavior on social networks.
Unfortunately, suicide is a terrible problem affecting many countries around the world. Among them is Canada. According to the Canadian Association for Mental Health, about 4,000 people take their own lives annually. In this country, suicide is the second most common cause of death among teens between the ages of 10 and 19but it is men between the ages of 40 and 50 who have the highest rate in the country; women, on the other hand, make suicide attempts 4 times more often than men.
Faced with these sad facts, the Canadian government plans to use artificial intelligence to reverse this trend and limit this social problem. As we know, there are already algorithms capable of detecting suicidal tendencies, and even machine learning able to predict symptoms of depression. On this occasion, the task will fall to Advanced Symbols, a signature of artificial intelligence based in Ottawa. The company’s goal is to be able to predict which areas of Canada could suffer an increase in suicidal behavior.
Advanced Symbolics technology is called Polly. Created in 2012, Polly harnesses the power of social media to forecast geographic peaks in suicidal behavior. This tool analyzes Canadian social media user posts to learn more about suicide-related behavioral patterns. Polly’s knowledge and intelligence continuously evolves and teaches herself; his job is not to point out particular individuals with suicidal tendencies, but to identify, through the analysis of the information obtained in social networks, places in Canada where there is an abnormal percentage of population with suicidal tendencies.
Detect suicidal behavior on social networks
“We use private techniques that are part of the industry standard. The same ones that are used by Statistics Canada and all other organizations work and handle personal information, ” he explains Kenton White, chief scientist at Advanced Symbolics. With this artificial intelligence, the Canadian government will be able to develop an action plan, as well as research on suicidal tendencies in the population, based on suicidal behaviors detected and identified by Polly on social networks. In this way, new strategies and prevention programs based on online data associated with users who present suicidal behaviors can be studied and thus prevent a crisis from erupting, all without identifying the individual.
“We don’t identify anyone, we don’t look for anyone, we don’t want to know everything about a particular individual. We just want to know as much as we can about the behavior of a population of more than 160,000 Canadian social accounts,” says White. When the trend is identified and it is determined that there is a real risk, the company informs the government so that it can allocate adequate resources in the designated area.
The Canadian government has stated that the pilot project will last three months. After this period, it will have to be determined whether or not this continuous monitoring will be necessary to combat this serious problem.
