4 ms·
Somewhat trusting person than I am, I can see this being an initiative borne out of genuine sympathy without sinister motive...but I can see a few potential pit
by danso 10y ago
Somewhat trusting person than I am, I can see this being an initiative borne out of genuine sympathy without sinister motive...but I can see a few potential pitfalls that, depending on how they're handled, could lead to unintended consequences.
One of the more obvious obstacles will just be noise. The ideal situation is someone who posts a clearly suicidal note late at night and a single friend sees it in time to act...this is the kind of signal that a huge network like FB can uniquely catch because of its size.
But what about all the more ambiguous situations? A friend with hundreds of FB friends has gone through a horrible breakup and other real life failures, and he starts posting notes that are potentially signs of suicidal thought...or, at least a dozen of his hundreds of friends think so, and they all flag it as suicidal. What should FB do then? At least a good number of these incidents will be false positives and people will start complaining about FB's human-backed algorithm for discerning potentially suicidal users.
The imminent danger is that someone will post notes that are suicidal in retrospect, people will flag it, and the user will commit suicide anyway. It won't be FB's fault...after all, AFAIK, they're just offering to send a friendly help screen to the user, not call the cops. But that won't be how the media or the deceased user's friends will see it, and the revelation that FB's powerful network has a human/algorithm hybrid element will be controversial in the same way that the trending news kerfluffle was. Perhaps there will be demands that FB should have done more, and that cops should be involved in the decision making...
But the seemingly inevitable decision would seem to be: why even bother waiting for humans to flag suicidal behavior? For a user who has a history of regular activity on the service, I'm sure there's a general profile they fit in terms of sentiment of messages and reduced engagement on the service, just like FB is reportedly able to tell how soon given couple might break up. Here's no reason why FB couldn't augment a profile with a hidden "depression" index to help the suicide response team triage reports...but the existence of such a calculation will undoubtedly be controversial.
And then there's the question of: if FB can calculate such a depression index, why wait for it to reach a high level before taking action? And, why should action be limited to sending an explicit anti-suicide message? Shouldn't FB be obligated to populate the user's news feed with happier messages, more reminders that they have a family, fewer reminders that their former lovers have found new lovers, etc?
And of course, for more controversy, we could always replace "suicide" with other kinds of unwanted behavior.
- forgottenpass 10y agoAnd of course, for more controversy, we could always replace "suicide" with other kinds of unwanted behavior. I can't shake the feeling that social networks are sleepwalking towards the very primitive versions of whatever the human analogue of a Skinner box is. I know gamification is a thing. Along with the other types of psychological manipulation in games. But this feels different to me. I get that suicide prevention is a prima facie positive, but I wonder what Facebook's ethics board has to say about this type of behavior in general. I guess in 20 years it will probably be just like advertising. We know it's inherently manipulative, but try to ignore what we know to be it's goals. And lie to ourselves and eachother about the extent to which it can alter our behavior.