Artificial Intelligence Could Prevent Accidental Selfie Deaths


People have fallen off buildings, drowned, crashed planes and even shot themselves while trying to snap the perfect selfie.

Now computer experts have analyzed selfie deaths and may have a solution — an app that could warn people if they were in a death-by-selfie zone.

Courtesy: Flickr
Courtesy: Flickr

“We found that most common reason of selfie death was height-related. These involve people falling off buildings or mountains while trying to take dangerous selfies,” Hemank Lamba of Carnegie Mellon University, Ponnurangam Kumaraguru of the Indraprastha Institute of Information Technology and colleagues wrote in their report.

The research team began tracking selfie-related deaths since March 2014, and found that drowning and being hit by trains came in at a close second for selfie deaths. They also discovered that most selfie deaths occurred in India (76), followed by Pakistan (9), the U.S. (8) and Russia (6).

“There was a news article that was circulated in my research group about a death by selfie during summer 2016,” Ponnurangam Kumaraguru, a professor at Indraprastha Institute of Information Technology in Delhi, told Digital Trends. “I was disturbed by reading it, exchanged some emails on this topic and found little work — especially from a technology standpoint — [had been carried out]. Our group is always interested in working on topics, technologies, solutions and systems that have real-world impact, so we jumped on it.”

Kumaraguru and his team went through news reports on selfie-related deaths and used data mining techniques to figure out exactly where and when people were in the most danger to die from a selfie mishap. Their research included a man from Washington, who shot himself while taking a selfie with a gun — he allegedly thought that gun wasn’t loaded.

Courtesy: Flickr

They then programmed an artificial intelligence (AI) system to go through selfies on Twitter, and then it gave a prediction of whether the selfie was dangerous or not. The AI system calculated factors like how close the selfie was to train tracks or from what height it was being taken. According to the published paper, the AI system had a 73 percent accuracy when analyzing how dangerous a selfie was.

“One of the directions that we are working on is to have the camera give the user information about [whether or not a particular location is] dangerous, with some score attached to it,” Kumaraguru continued to tell Digital Trends.

Currently, Kumaraguru and his team are working on developing software that could warn users of potential selfie-taking dangers, or even turn the camera function off if the AI deems the selfie too risky.

Danielle Tarasiuk

Danielle Tarasiuk is a multimedia journalist based in Los Angeles. Her work has been published on, Yahoo! Sports, KCET, and NPR-affiliate stations KPCC and KCRW. She’s a proud Sarah Lawrence College and USC Annenberg alumn.