Facebook activated its “safety check” feature in Lahore, Pakistan, on Sunday, after a suicide bomber killed more than 70 people who had gathered in a city park to celebrate Easter. When “safety check” is on, Facebook users in the immediate area of a disaster or an attack are alerted and asked to “check in” and let people know they’re safe. But on Sunday, something went wrong.

A “bug,” Facebook said in a post on Sunday, mistakenly sent messages to Facebook users who were nowhere near Lahore at the time of the attack.

According to social media postings from people who said they receive these notices in error, versions of the “safety check” notice seen above went out to quite a few people in the United States and Britain. In the comment thread of Facebook’s note about the bug, there are reports of messages going to Australia and the Democratic Republic of Congo. It’s not clear right now exactly how widespread the error was, however.

A reporter from the Toronto Star noted that the text version of the safety check feature mentioned an “explosion,” but didn’t specify exactly where, making the message seem just a bit more alarming:

“Unfortunately, many people not affected by the crisis received a notification asking if they were okay.” Facebook’s statement about the errant messages reads.  “This kind of bug is counter to our intent. We worked quickly to resolve the issue and we apologize to anyone who mistakenly received the notification.”

Facebook has only recently begun to use its safety check feature during conflict situations and terrorist attacks — the first time it did so was during the November attacks in Paris. At the time, Facebook said that it would start activating the tool more frequently, following widespread questions about how the social media company decides when to use it.

Since the start of 2016, Safety check has been activated in response to natural disasters in Brazil, Fiji, Taiwan, and India; and after the bombings in Turkey and Brussels.

Liked that? Read more: