Monitoring Social Media Could Help Suicidal Veterans
The Geisel School of Medicine is helping the Durkheim Project for veterans.
CREDIT: Durkheim Project
Can people's tweets and status updates indicate that they are about to hurt or kill themselves?
That's the premise of the Durkheim Project, an initiative launched in collaboration with the Geisel School of Medicine at Dartmouth College. The project will use artificial intelligence to monitor war veterans' social media activities in an effort to prevent them from committing suicide or inflicting self-harm. If the research goes well, the technology could eventually protect anyone who chooses to enroll.
According to a 2012 report from the U.S. Department of Veterans Affairs, about 14.3 percent of vets commit suicide. The national average for all Americans is well below 1 percent.
Chris Poulin, the project's principal investigator, hopes that the Durkheim Project — named for French sociologist and pioneer in suicide research Émile Durkheim — can use key words and phrases that veterans post on Facebook, Twitter, LinkedIn and other social platforms to help identify veterans at risk of inflicting self-harm. This isn't snooping. Veterans must opt in to utilize the project's apps, which monitor activity both on computers and mobile devices.
Not just gripes
Although confidentiality concerns prevented Poulin from sharing specific words and phrases that the Durkheim Project seeks out, he explained that there was a very clear difference between suicidal tendencies and the general grievances many people enjoy broadcasting on social media.
"I can tell you that the general phrases in the suicidal cohort … relate to a sense of inward agitation," Poulin said, citing the project's initial data-mining research, which examined a huge database of veterans, some of whom were mentally sound, and some of whom went on to commit suicide.
"Words related to agitation and inward blame were coming up more frequently. People who are suicidal only complain about a very small amount of things," he said. "The [control] group was complaining about many things that the suicidal group did not complain about." In fact, Poulin said that every key word or phrase in the suicidal cohort could be traced back to established medical literature on suicidal behavior.
Because the current phase of the Durkheim Project doesn't include intervening if someone appears to be at risk, it is possible that some veterans involved may go on to commit suicide. "It's a morbid side of the research; I'm not going to deny that," Poulin said. "I don't know if it'll be our best data, but it would certainly be informative if we found out that someone took their life." [See also: 10 Ways to Protect Yourself on Social Media Websites]
This approach may seem cold, but Poulin points out that researching mortality rates is standard procedure in any medical study that involves life-threatening circumstances. "In cancer treatment, [you study] if someone survived a cancer drug or if they didn't … It informs your study. It's a depressing side of the research, but it's not our primary intent."
If the research progresses well among veterans, the Durkheim Project could eventually become a tool for everyday mental-health professionals and their patients. "[They could] use it as a clinical tool, like a cholesterol test," Poulin said.
But the project has a possible caveat: The people who might benefit most from the project may decide not to enroll in it. "It would have to work like any other treatment or screening program," Poulin said. "A doctor could recommend it, but you and I are free to reject the doctor's recommendation … Many of the people we could help are not going to opt in. I do expect that."
Even though it may not help every single veteran at risk of committing suicide, at least it's a start in addressing the epidemic of suicide among those who served.