That’s what happened on Wednesday, when a group of four men and women in Chicago inflicted a horrific attack on a young man who was bound, gagged, and cut with a knife by one of the assailants. According to police, he eventually escaped and was later taken to the hospital. The four were arrested and charged with kidnapping, battery, and hate crimes because the suspects are black and the victim is white, and because racial slurs were used during the attack.

At one point the video stream—which has since been removed—had more than 16,000 viewers, according to a number of observers who watched it. The Guardian reported that several Facebook users interacted with the group while they were abusing the boy (a classmate of one of the assailants), posting comments that the attackers then responded to on camera.

Get Data Sheet, Fortune’s technology newsletter.

This is probably not what Facebook had in mind when it launched Facebook Live last year, or when a Facebook executive said that within five years, the Facebook news feed will be all video. But similar kinds of violence have become commonplace on the platform.

Last year, the death of Antonio Perkins was streamed live on Facebook after he was shot in a drive-by attack in Chicago while filming himself drinking with friends on the sidewalk. And a second man in Chicago was shot at while streaming on Facebook Live, although he survived the attack. Brian “Sugar Ray” Fields was recording video while outside a store when a gunman walked up and shot at him.

In almost every case, including the most recent one, Facebook was criticized for leaving the video stream up for so long. And yet, often when it removes video streams or photos or clips, the social network gets criticized (justifiably, in many cases) for engaging in censorship.

When Philando Castile was shot by police during a routine roadside stop in Minneapolis, and his death was filmed by his girlfriend Lavish Reynolds as her daughter sat in the back seat, the footage helped galvanize protests about police violence against blacks in the U.S.

In that sense, even a violent video stream can have positive effects. But Facebook would probably much rather that Facebook Live was about birthdays, weddings, and new puppies—and if it wants to eventually introduce video advertising (which it does) having disturbing content isn’t a great way to attract brands to your platform. Just ask Twitter.

This is how modern political advertising works:

In many ways, these Live video cases help emphasize how much Facebook continues to struggle with the challenges of being a reluctant media company. As with the “fake news” dilemma, it would very much like to just be a neutral platform, a distributor of content—but it can’t help being drawn into these ethical discussions because of how it is being used.

It’s not alone in facing these kinds of dilemmas, of course. YouTube has faced similar challenges whenever there is a video clip of a beheading—something that is also arguably a news event worth disseminating—and so has Twitter.

In Twitter’s case, and to some extent YouTube’s as well, the company has made a commitment to free speech, and to distributing content regardless of whether it is disturbing (although in Twitter’s case that seems to be weakening). Facebook has never really grappled with those questions in any kind of public way, perhaps because it is afraid of the implications if it embraces its nature as a media outlet.

Until it does so, these ethical dilemmas will continue to crop up. And the company will likely continue to be criticized for not having a coherent way of dealing with them. And it doesn’t sound like Facebook’s new “head of news” will help the social network figure out its role in this evolving media landscape—the company announced Friday that it has hired former NBC host Campbell Brown to be its head of news partnerships, but said she will not be involved in content decisions.