Facebook’s marketing for its Facebook Live video streaming service focuses on the potential for sharing intimate moments from its users’ lives like weddings and birthdays. But its video platform has also become a popular way to share much darker moments.
While many families were enjoying an Easter or Passover meal on Sunday, a man in Cleveland was uploading video of himself shooting and killing a 74-year-old grandfather. After he was done, the killer logged on to the social network to talk about his actions.
The killing is the latest in a series of incidents that have cast a shadow over Facebook’s live video streaming service. Although the murder was not actually live-streamed (contrary to some initial news reports), it still raises significant questions about Facebook’s responsibilities in such cases, since the video remained available for hours after the shooting.
The social network released a statement to CNN late Sunday saying “this is a horrific crime and we do not allow this kind of content.” But for many, the damage had already been done. And the fact that the killer spent time after the murder live-streaming video of himself talking about the incident seemed to add insult to injury.
Facebook also posted a statement late Monday saying the “terrible series of events” had no place on the platform and “goes against our policies and everything we stand for.” It said it took a long time for the video to be removed because users didn’t flag it as offensive for over an hour. “We know we need to do better,” the post added.
Since Facebook introduced its live-streaming video service a year ago, there have been several cases of people broadcasting deaths and other violent acts to the network’s billions of users. In some cases, the company has taken swift action to remove the videos, but in others it has chosen to leave them up with a warning about the disturbing content.
Last year, for example, Antonio Perkins of Chicago was shot and killed in a drive-by attack while he was live-streaming himself drinking with friends on the sidewalk in a residential neighborhood. The video was watched hundreds of thousands of times within a matter of hours.
Despite the violence of the attack, and the fact that Perkins died as a result, Facebook told CNN at the time that the clip was left up because it didn’t violate the company’s standards.
In July, Philando Castile of Minneapolis was shot by police during a routine roadside stop, and his death was filmed and live-streamed by his girlfriend Lavish Reynolds, while her young daughter sat in the back seat. In that case, the footage—which Facebook didn’t remove until much later—helped galvanize protests about police violence against blacks in the U.S.
Get Data Sheet, Fortune’s technology newsletter.
There have also been several cases that were equally disturbing, even if they didn’t involve murder. In January, a 12-year-old girl uploaded a live-stream of her own suicide to Facebook, in a clip that showed her tying a rope around a tree and hanging herself. Facebook was criticized for leaving the video clip up for more than two weeks.
There were also reports this past weekend that a man in India live-streamed his own suicide using the Facebook Live feature.
In January, a group of four men and women in Chicago live-streamed an attack on a young developmentally-disabled man who was bound, gagged, and cut with a knife (he later escaped, and the four were eventually arrested by police).
At one point, the video stream had more than 16,000 simultaneous viewers, and several Facebook users interacted with the attackers while they were abusing the boy. According to a number of reports, they posted comments that the attackers then responded to on camera.
Just last month, several teenage boys in Chicago live-streamed the sexual assault of a 15-year-old girl using Facebook. According to police, at one point more than 40 people were watching the attack, but no one called 911 or contacted the authorities.
Facebook isn’t the only company that has to deal with this kind of violent content— similar clips also get uploaded to YouTube, including the ones of the 12-year-old’s suicide and the latest shooting. But Facebook has spent so much time and energy promoting its live feature that it has become a lightning rod for criticism in such cases.
And whatever the social network may think of the ethics of live-streaming deaths or leaving clips up for weeks, having newspaper and TV headlines using the term “Facebook murder” or “Facebook killing” probably isn’t something the company wants to deal with.
If it wanted to, Facebook could restrict live-streaming in a variety of ways (to only verified users, for example) or it could get rid of the feature completely. But it’s unlikely to do either of those things, because it sees video as a huge part of its future. And in the meantime, we are likely to see examples of its darker uses crop up more and more frequently.