Some 70 percent of youth aged 13–17 see real-world violence on social media, according to a 2024 Youth Endowment Fund survey of teens in England and Wales. TikTok users reported the highest exposure at 44 percent of respondents in the preceding 12 months. X users followed with 43 percent, though less than a quarter of respondents reported using the platform. Facebook, Snapchat, and Instagram weren’t far behind, with exposure rates of 31–33 percent.

The recent assassination of conservative commentator Charlie Kirk was “sadly, an interesting case study,” Blevins said. The extremely graphic videos of the incident were seen by millions of people on social media despite platform restrictions.

Too Much

Research literature has talked for years about the issue of “desensitization” by repeated exposure to violence in digital media.

In real life, Americans overall have been fortunate enough to only sporadically experience violence on a personal level, Selepak said.

“We don’t experience violence every day, multiple times a day, in the real world; but on social media, especially because people are spending hours each day [on those platforms], you can be exposed to it for hours each day repeatedly,” he said.

“Outside of people engaged in war, there’s never been an equivalent to this in human history.”

Even those with the most real-life exposure to violence, such as first responders, often see only its aftermath, which leads them to appreciate its destructive consequences.

Online, on the other hand, a person can easily get pulled into watching a near-endless stream of real-world violence detached from its consequences, Selepak noted.

A major issue is the use of algorithms to personalize content for each user, the experts agreed. Once a person stumbles upon a few videos of a certain kind, the algorithm picks up on the apparent preference and supplies more. Even if the platform restricts graphic violence, there’s plenty of borderline content that makes it through, yet is still brutal—and it’s exactly the borderline content that often goes viral.

“Content that does anger us, that does upset us, that is upsetting content, we’ll spend more time consuming that, commenting on it, engaging with it, which leads to that content, based on the algorithm, then being more visible to more people, because it’s getting so much engagement,” Selepak said.

And restricting such content would hurt the platforms’ bottom line, he pointed out.

“The platforms make money by how much time we spend on there, seeing ads.”

Further complicating the issue is that violent content has real information value in some contexts.

Algorithms, however, lack the ability to make such distinctions because they lack human sensibility to begin with.

“Algorithms are really good with content, but certainly not context and ethics,” Blevins said. “They don’t necessarily come with that moral compass.”

Selepak concurred.

“The algorithm doesn’t necessarily care what we’re consuming. It just wants us to consume,” he said.

Conflict Magnifier

Social media doesn’t just expose people to violence; it also facilitates it.

“Adolescents often do not go online with the intention to fight. Rather, as they are acutely aware, social media is a unique environment that amplifies and changes experience of conflict,” it said.

Calls for Violence

Beyond visual gore, social media also provides a forum for violence advocacy. The experts refrained from singling out any particular political group, but highlighted the significant volume of such content, despite various platform restrictions.

“If you’re interested in a particular ideology or a particular group or a particular phenomenon, start a fresh account … feed in those keywords, and just watch what the algorithm feeds you,” Blevins said. “It could be eye-popping.”

With algorithms constantly recommending more similar content, people using those platforms can easily gain the impression that an extreme or fringe position is normal and commonplace, the experts warned.

There’s also a new phenomenon of “sycophantic” personal AI chatbots that affirm the user’s beliefs to an unhealthy degree, Blevins noted.

Then there’s content that doesn’t call for violence outright, but provides indirect justification for it.

A man holds a phone displaying the TikTok app.

“If you’re on Meta or YouTube, X as well as TikTok, there’s this constant drumbeat of one side is fascist and Nazis and what we would consider the absolute worst of the worst of humankind,” Selepak said.
“That kind of dehumanization, and the repeated presentation of that dehumanization, can make violence or accepting violence against that group much more possible.”

With a growing body of research showing the negative effects of social media, at least in its current form, public opinion toward it has been shifting as well.