This App Prevents Thrill-Seeking Selfie Takers From Dying (Hopefully) Telling someone it's dangerous when they're already hanging off a skyscraper may not help save many lives.
By Matthew Humphries •
This story originally appeared on PCMag
When Pokémon Go launched in July, it had millions of us walking the streets hunting for pocket monsters. But it also caused a number of accidents and even deaths because people were looking at their phones rather than the world around them. The Pokémon Go craze has died down, but we're facing another, longer-term smartphone-related problem: selfie deaths.
It may sound ridiculous, but the number of deaths occurring while attempting to capture a selfie is increasing every year. According to the BBC, there were 15 selfie deaths in 2014, 39 in 2015 and 73 in 2016 as of August. It seems likely that figure will be closing in on 200 for 2017 unless something is done, but a PhD student by the name of Hemank Lambda is on the case.
Lambda is studying software research in the School of Computer Science at Carnegie Mellon University. Along with a team of friends, he's developing an app that keeps selfie takers safe by assessing danger and warning them as they line up their next (and hopefully not last) selfie.
Studying previous selfie deaths revealed most involved falling from a great height; off a cliff, for example. In India (where most selfie deaths happen), the majority of deaths are linked to trains; in the U.S. and Russia, it's guns. Each area of the world has its own dangerous past times, and they clearly become more dangerous when you try to capture your face while doing them.
Lambda's app will analyze the situation when it detects a selfie is imminent and respond accordingly. If it detects a dangerous situation -- perhaps the handset is at a great height or it recognizes a train moving in the background -- a warning will be issued. Sensors on the phone, GPS coordinates, and computer-vision techniques will all be used to gather as much information as possible so as to better assess risk.
So far, Lambda's app has been tested on 3,000 selfies and manages to assess risk correctly 70 percent of the time. I suspect even if his team achieves 100 percent accuracy, thrill seekers will ignore the warnings, and that's if they can even see them while hanging off a cliff face ... in high wind ... with one hand.
