Official newspaper of The University of Texas at Austin

The Daily Texan

Official newspaper of The University of Texas at Austin

The Daily Texan

Official newspaper of The University of Texas at Austin

The Daily Texan

Advertise in our classifieds section
Your classified listing could be here!
October 4, 2022
LISTEN IN

Facebook puts safety in the hands of their users

Facebook wants to make crisis-reporting a community effort.

On Wednesday, Peter Cottle, a founder and lead engineer of Facebook’s Safety Check feature, presented a talk on the addition of crowdsourcing to Facebook’s existing crisis-reporting software. The UT computer science department hosted and sponsored the talk, which was held in the Gates Dell Complex.

According to Facebook, Safety Check currently allows users to quickly notify friends and family that they are safe following a natural disaster or other tragedy. 


Safety Check began during a 72-hour hackathon that Cottle participated in after he transitioned from an internship to full-time engineer at Facebook in 2013. According to Cottle, Safety Check was inspired by the behaviors of Facebook users during the Boston Marathon.

“We started to look into what users were doing during disasters, and they were broadcasting safety by making posts,” Cottle said. “Posting on your wall is a nice way to let everyone know you’re okay, but it isn’t the best product experience.”

According to Cottle, the major problem with this is that there is no guarantee a post will be seen by people, and users tend to avoid making posts even in disaster situations. In fact, only 4 percent of users who open up Facebook in disaster areas post anything.

“We thought clicking a button would be a great alternative,” Cottle said. 

According to Cottle, with access to users, their locations and the people who care about them, Facebook was perfectly positioned to provide this service.

The first highly trafficked launch of Safety Check coincided with the 7.8-magnitude earthquake in Nepal, which left more than 8,000 people dead.

“Before this, no one really knew what the tool was,” Cottle said. “This was the first time users expected Safety Check to be turned on.”

Safety Check expanded beyond natural disasters with their second launch: the Paris terror attacks.

But Safety Check’s widespread use did not come without a slew of major obstacles, according to Cottle.

The first issues were technical. Keeping track of Facebook’s 1.7 billion users at all times would require an unrealistic amount of memory, and it would take too long to find people in the affected area. Second, designating disaster areas around the world is difficult to do because of bias in the news and the unreliability of user data.

Another problem was that the world began to rely on Facebook to designate disaster status.

“The top comments were consistently ‘Why didn’t you turn it on here or there?’” Cottle said. “This was a tough time because Safety Check became this barometer for how significant an event was, and Facebook didn’t want to further reinforce Western bias of news coverage.”

To address these problems, Cottle and his engineers decided it was time to put crisis reporting in the hands of the people on the ground.

Once Facebook learns when and where disasters have occurred using police scanners and user activity, users in the area take over by simply opening their Facebook news feed.

“What we ended up actually doing is when we launch Safety Check during a disaster, we ‘hook’ on to anyone who opens their news feed,” Cottle said. “So anytime anyone opens Facebook, a small piece of our code executes. First, we check if they are in the affected area. If they aren’t, we bail. If they are, we activate a Safety Check for them, and we start checking if their friends are in the affected area. If they are, we send them a notification and so on.”

According to Cottle, once one user in the disaster zone opens their news feed, every Facebook user in the affected area can be found within 10 minutes. For Nepal, within those 10 minutes, they checked in 1 billion users, and invited 3 million to Safety Check. 

“This method biases towards active [users] who are really connected, so you get people who are likely to use Safety Check [that] check on friends and will engage in the process,” Cottle said. 

For other areas, the algorithm scales the computing power based on the size of an area. This allows both small crises and high-density disaster areas such as war zones to be addressed automatically, eliminating any possible bias. 

Before crowdsourcing was implemented in June 2016, Facebook was responsible for identifying crisis areas and sending notifications to everyone in those areas. Now, every Safety Check launch and subsequent spread relies on Facebook activity by the people on the ground who will benefit from it. 

UT computer science sophomore Joel Swiatek attended the talk and wondered how Safety Check influences Facebook’s current and future relationship with emergency services.

“Do you see first responders interacting with Safety Check in the future?” Swiatek asked of Cottle.

Cottle said that while in the future Safety Check would like to help first responders handle disasters more efficiently, there are many legal issues surrounding this relationship. Regardless of these issues, Cottle said Safety Check will continue to make a difference in crisis response.

“Since Safety Check started, we had about 28 launches. Since June, we have had 232 launches automatically and algorithmically driven,” Cottle said. “No one has done filtered worldwide crisis chatter, but because of Facebook’s 1.7 billion user reach, we are doing it.” 

More to Discover
Activate Search
Facebook puts safety in the hands of their users