AI to be used to predict outbreaks of suicide
January 10, 2018
OTTAWA – The Canadian government will soon hire an Ottawa-based company specializing in social media monitoring and artificial intelligence to forecast potential spikes in suicide risk. A contract with Advanced Symbolics Inc., an AI and market research firm, is set to be finalized next month.
CBC News reported that a pilot project will be conducted for three months, after which the government “will determine if future work would be useful for ongoing suicide surveillance.” Suicide is the second-leading cause of death for Canadians aged 10 to 19, according to the Public Health Agency of Canada.
“To help prevent suicide, develop effective prevention programs and recognize ways to intervene earlier, we must first understand the various patterns and characteristics of suicide-related behaviours,” a PHAC spokesperson said in an email statement.
“PHAC is exploring ways to pilot a new approach to assist in identifying patterns, based on online data, associated with users who discuss suicide-related behaviours.”
Instead of calling people to assess public opinion, Advanced Symbolics conducts its market research by identifying and tracking social media accounts to build a representative sample of a population.
Many phone surveys poll roughly 1,500 people, but Advanced Symbolics said its representative sample of Canada’s population uses more than 160,000 social media accounts. And the company said its market research method has been accurate where many others have failed.
“We’re the only research firm in the world that was able to accurately predict Brexit, the Hillary and Trump election, and the Canadian election of 2015,” said CEO Erin Kelly.
Advanced Symbolics said its artificial intelligence looks for trends, not individual cases.
“It would be a bit freaky if we built something that monitors what everyone is saying and then the government contacts you and says, ‘Hi, our computer AI has said we think you’re likely to kill yourself’,” said Kenton White (pictured), chief scientist with Advanced Symbolics.
Instead, the AI will flag communities or regions where multiple suicides could be likely. For example, Cape Breton Island was left reeling last year after three teenagers in the region died by suicide.
“The spike that happened in Cape Breton, as unfortunate as it is, we can learn patterns from that,” White said. “We can also learn patterns from what happened in Saskatchewan, patterns from Northern communities, patterns from college students.”
“We’re not violating anybody’s privacy – it’s all public posts,” added Kelly. “We create representative samples of populations on social media, and we observe their behaviour without disturbing it.”
In the weeks following the suicides in Cape Breton, the provincial government sent additional counsellors and mental health experts to the region.
According to Advanced Symbolics, artificial intelligence could offer a two- to three-month warning before a potential spike in suicides occurs. And sharing that information with government officials could prompt them to mobilize mental health resources before a crisis, instead of afterward.
The company will begin defining suicide-related behaviour in January, with monitoring slated to start later in 2018.