This study seeks to identify a recognizable pattern in social network status updates of a potentially suicidal person.
Research-wise, it’s a solid premise. Human behavior is predictable in a lot of ways; For many decades, experts have recognized common “signs” that a person is thinking about self-harm. It turns out that we humans think in similar ways. We probably tweet and post FB updates in similar ways, too.
One of the more interesting points with this article, though, is the comments. Lots of talk there about spying and the observer-expectancy effect (where the results are skewed because the subjects know they are being watched). People seem a bit nervous about their statuses being analyzed, even for the sake of saving lives.
But what if no human saw the data? I can see a scenario where the computer sends a human-written note saying “Your statuses indicate you’ve been struggling. Here are some helplines you may find useful.” No privacy invasion. Just a simple, not-personal algorithm doing its job.
When life insurance companies are on your digital trail, however, that could be a devastating thing. I was interviewed for life insurance when I was in my young 30’s. The agent was so very rude and insistent about assessing my mental health history for suicide risk that I had to ask for a manager. (The manager stopped the questioning and put the policy through immediately. I was a prime, healthy candidate for them and she knew it).
Again, this is the brave(ish), new(ish) world, a place where computers might care for you more than humans do.