Checking the Safety Check

April 11, 2017


When the attacks in London earlier this month happened, many people used Facebook’s “Safety Check” functions to alert friends and family that they were safe.

When launching the feature, Facebook CEO Mark Zuckerberg announced, “When disasters happen, people need to know their loved ones are safe. It’s moments like this that being able to connect really matters.” As Wired explains, “When activated, Safety Check locates Facebook users near a disaster site through the city they list on their profile, or from where they last used the Internet. Users then receive a notification asking to confirm that they’re safe or to say that they weren’t in the affected area. Those who choose ‘safe’ generate a notification to their friends and followers, who can track how many of their friends were affected.”

I’m sure the effort is well intended, but my natural skepticism led me to wonder just how useful it really is. There are about 10 million people in London at any given time (8.5 million residents plus another 1.5 million visitors per month, roughly) and the chances that any given one of them will be harmed or killed in terrorism or a natural disaster is very remote.

To be clear, the feature likely gives comfort to the small percentage of potentially affected people marked safe, and that’s worth something. But someone not checking in as Safe doesn’t mean anything: They could be a continent away and oblivious to the event, or they could be dead. Statistically, it’s likely that over 99% of the people who don’t check in are safe as well. People may not check in as “Safe” for any number of reasons. Perhaps because they don’t use the function, don’t think to do it amid the chaos, are unaware that anyone would think they’d be in harm’s way, are in contact with friends and family anyway, prefer to protect their privacy, and so on.

For example I know a handful of people who live in London (and I probably know friends and family who happened to be London for business or pleasure at the time and I didn’t know it), but I didn’t think to check to see if they marked Safe or not, nor would I have, since the odds they were affected are millions to one.

Numeracy and a good grasp of statistical probability can provide some comfort, but the Facebook function may in fact make matters worse. As Guardian columnist Tim Burrows noted, “The attack on the Palace of Westminster was grave and tragic, but, as rapidly became apparent, it was also crudely undertaken and successfully contained by the police and emergency services. Four people died during the atrocity, 29 people were injured. The need for a vast network of people to check in as safe when only a small fraction of the city was in peril doesn’t stand up to much scrutiny. The problem is one of scale. On the mobile version of Facebook’s Safety Check page, the ‘affected area’ shows a map encompassing as far north as Watford, Hertfordshire, as far west as Slough, Berkshire and as far east as Basildon, Essex. Such an area is more suited to be ‘affected’ by a thermonuclear weapon than such a crude attack in Westminster.”

Another reason to be skeptical of the utility of Safety Check is that news media are very quick to identify the victims anyway in the course of reporting the story. This is especially true when there is a relatively small number of victims. In the recent Westminster London attack, for example, four people were killed and their identities (along with identifying details such as nationality and occupation) were broadcast widely on news and social media. Early reports identified three of the four victims as British nationals, and the fourth an American. This information alone rules out a huge number of potential victims; they were not Canadian, nor Brazilian or Chinese, and so on. Of course one may have friends and family members of any nationality, but the point is that even the smallest pieces of information can provide a lot of information about who the victims might be. Just watching a few minutes of the news will give you a pretty good idea of whether your loved ones have become victims.

In May 2015 following the Nepal earthquake, an article on Buzzfeed provoked outrage when it reported that people in America and the United Kingdom, who were seemingly not affected by the disaster, were marking themselves Safe. This was interpreted as a tasteless prank, though Snopes later clarified that it was due at least in part to software glitches, and that “While it’s true that some users inaccurately marked themselves as safe following the Nepal quake, it’s not clear such actions were undertaken maliciously or disrespectfully.”

As The Atlantic noted, the April 2015 devastating earthquake in Nepal “revealed some of the limits to Safety Check. Smartphone penetration in the country-one of Asia’s poorest-is low, and six Nepalese out of seven are not registered on the social network. Electricity in the country is unreliable even during normal times, and there were reports of extensive power outages throughout Kathmandu in the hours after the quake.” Two years later, other recent mass-casualty disasters reinforce this flaw. Last month a landslide swept through an enormous garbage dump outside of the Ethiopian capital of Addis Ababa killing at least 100 people and leaving dozens of others missing and homeless. And last week at least 260 people were killed in a mudslide in rural Colombia following torrential rain, with hundreds more injured and missing.

These and other similar disasters highlight the not-so-hidden (but inevitable) class disparity inherent in the Safety Check mechanism, one that might subtly suggest that privileged lives are more important than those of the poor, who may not have the money or time to check in on social media–or may have no internet access at all. In the end, Guardian columnist Burrows notes that in the recent London attacks Safety Check “spread unnecessary worry about people who were simply nowhere near the affected area. The big question for Facebook is: when can a Safety Check make people feel less safe?”