First it was the nightmarish stabbing of Iryna Zarutska, a 23-year-old Ukrainian refugee, as she sat on a train in Charlotte, N.C., minding her own business. Then it was the horrifying shooting of Charlie Kirk, the 31-year-old conservative activist, as he addressed a group of students at Utah Valley University. Both struck terror in countless Americans fearful for their own safety and for the safety of our public spaces and our democracy.
The tragedies had something else in common, though: They both generated extremely graphic videos of the victims’ last moments, detailed enough to show the second that metal struck flesh and wrought it awful damage. Since then, shared by many and further amplified by digital algorithms that favor intense emotions, these videos have been endlessly replayed across social media. Countless users have commented on them, zoomed in on them, slowed them to a crawl, theorized about them or marked them up with arrows and diagrams and published the results. Ad nauseam.
In the nascent stages of social media, I was an optimist about unfiltered imagery. I thought, as did others, that unfiltered images from news events might make people more empathetic toward victims of natural disasters, repression or systemic violence. I also hoped raw reality from conflict zones would challenge the sanitized, cinematic version of war that too many people held or might force them to care about conflicts they were otherwise happy to ignore.
That’s not what happened. Today there are more cameras than ever, and we’re drowning in videos documenting the last breaths of victim after victim. But instead of making us all more sensitive to the horrors that our fellow humans experience, instead of functioning as tools of understanding, graphic images like the videos of Zarutska and Kirk turn into something closer to viral snuff films when they are endlessly replayed. Reducing tragedy to voyeuristic content, they dehumanize not just the victim but all of us.
And as social media is woven ever deeper into every corner of our lives — school, work, civic engagement, religious participation — dehumanizing images like these become harder for even the squeamish to avoid. I asked my students about their experiences. One told me her father had casually encouraged her to watch the video of Kirk. At a restaurant where I was having dinner on Thursday, someone at the next table pulled out a phone and played the video for his companion.
And shortly after Kirk’s killing, social media got flooded by the next viral snuff film, a video of a gruesome beheading in Dallas. It sometimes seems there’s another shocking video every week.
Supplied with all this rich material, frenzied amateur sleuths role-play as if they were forensic scientists analyzing the Zapruder film. In the process, they often end up blaming the wrong people or propagating ever more absurd and harmful conspiracy theories. A number of people have been widely and wrongly identified as Kirk’s killer, endangering their lives and also probably impeding the investigation. And I’ve already seen widespread claims that Kirk isn’t dead at all and that the blood gushing from his neck was fake and triggered by a hidden mechanism. There, you can see it — or at least these self-appointed detectives announce that they can — if you zoom in on the video 1000X. So, rather than bringing the reality of gun violence home, these videos are helping at least some fraction of viewers deny a death that occurred in front of hundreds of witnesses.
The final moments of Iryna Zarutska, for example, were quickly deployed — because the suspect is Black and Zarutska was not — to advance a very specific argument about race or Democratic-run cities or liberal media bias. If you dug deep, you might have found reporting about the suspect’s history of violence and mental health struggles and his family’s desperate, fruitless efforts to get him adequate care or even to get him committed. It’s a problem all too many families know. But that part of the story was absent from the endless replays of Zarutska’s death, often with close-ups of her face right before she collapses. Instead of a sense of outrage leading to a search for a better solution for repeat violent offenders, her death generated calls for collective retribution and vigilante justice.
Obviously, interest in violent imagery is not something that started with or is limited to social media or digital technologies. The evening news programs that families used to gather around followed the maxim “If it bleeds, it leads” — violence first — which was good for ratings but bad for society. Studies have found that people who watch a lot of local news, in which crime is often sensationalized, tend to get a distorted view of how dangerous their neighborhood is. Viewers, especially older viewers, are more likely as a result to become anxious, even reclusive.
A gruesome wave of beheading videos by ISIS were many users’ first encounter with material of this kind. Pressured by the U.S. government, major platforms effectively kept the images from circulating. Nowadays you have to remember to turn off autoplay to avoid accidentally watching a beheading, as many people discovered after the grisly murder at a motel in Dallas.
Because cameras have become ubiquitous, the supply of violent images has multiplied many times over. Cultural and institutional barriers to disseminating them have decreased. Incentives for doing so also have multiplied. And U.S. government efforts to regulate any of it have all but disappeared. Kennedy assassination conspiracists had just a few frames of film to keep them busy for many decades. Now public events are often captured in many, many videos, recorded by closed-circuit TV or security cameras, audience or passerby phones, doorbell devices and more. And on social media there are no gatekeepers to decide how many viewings of a disturbing image is enough to meet public interest before tipping over into public harm. Nor is there much cultural decorum left to hold most people back from playing someone’s tragedy for the sake of their own audience engagement. The social media business model depends on it and rewards it.
Some mass shooters and assassins may depend on it, too. When they write manifestoes, wear clothing that bears a charged message or use certain weapons that have become symbolic of such shootings, are they producing content to be disseminated along with videos of their terrified victims’ last moments?
On Friday morning, we learned the identity of the person suspected of killing Kirk: a young white man in his early 20s, a former engineering student with strong high school grades. All the theories that the online sleuths had come up with, and all the supposed identifications of the shooter, appear to have been false. He was found not because of some crowdsourced forensic breakthrough, but because a family member reached out to a family friend who in turn reached out to law enforcement.
As for all the finger-pointing, all the hateful statements confidently accusing one group of people or another, I doubt corrections will be offered, and if they are, I doubt they will undo much of the damage.
Meanwhile millions, perhaps billions of people have watched and rewatched Kirk’s and Zarutska’s last moments as if they were video game clips or movie scenes instead of the dying moments of a man leaving behind young children or a young woman slain in the prime of her life. Virality achieved, but humanity — theirs and ours — lost.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
Zeynep Tufekci (@zeynep) is a professor of sociology and public affairs at Princeton University, the author of “Twitter and Tear Gas: The Power and Fragility of Networked Protest” and a New York Times Opinion columnist. @zeynep • Facebook
The post Social Media Reduced Two Horrific Killings to Cheap Snuff Films appeared first on New York Times.
