Facebook Mistakenly Send Safety Check Warnings About the Lahore Bombing Worldwide

BBC

Were you in Lahore on Sunday, or even in Pakistan? Well, even if the answer to both of those questions is 'no', Facebook may have decided that you were in danger of being caught in the blast radius of the suicide bombing which happened there on Sunday, regardless. The attack, which took place in a public park, claimed the lives of at least 72 people, leaving hundreds more critically injured. Many of the victims were children.

Thanks for the concern Facebook but Wrexham and Pakistan really aren't in the same area pic.twitter.com/VADIH4gudB

— cam (@camvallen) 27 March 2016

This brutal, abhorrent attack certainly warranted Facebook's 'Safety Check' feature, just perhaps not quite on the scale it ended up spanning. People as far flung as Australia, Europe and the Americas were asked to mark themselves safe, as Facebook had detected that they were dangerously close to the site of the explosion.

Facebook's errant "explosion" texts, which don't even specify "Pakistan," are even worse than the app/browser issue. pic.twitter.com/175rCOaIVM

— Daniel Dale (@ddale8) 27 March 2016

The real issue was that the messages didn't necessarily specify which 'explosion' it was on about, likely alarming, or even panicking the recipient. Facebook posted an official apology, citing the issue as a 'bug', but offering no further clarity beyond that.

We activated Safety Check today in Lahore, Pakistan, after an explosion that took place there. We hope the people in the...
Posted by Disaster Response on Facebook on Sunday, 27 March 2016

This isn't the first time the feature has landed Facebook in hot water. It was used during the Paris attacks, but not for the bombings in Beirut which had occurred the previous day. The argument they used at the time was that the feature was only effective for single, isolated incidents in its current form, rather than any kind of ongoing conflict. Even so, they pledged to be more even-handed with it going forward.

The biggest lesson to be learned from this is that the safety check should always specify the location in the message, regardless of whether it's sent via text, notification or whatever else. It's hard to say whether the bug had any negative impact on the feature's use in the actual danger area, but it seems unlikely, since anyone mistakenly marked as nearby would have to have been either very stupid or kind of evil to then mark themselves safe (ie nearby) and throw the whole thing out of whack.