The Flawed Algorithm: Dead Dog TikTok as an SEO Friendly Headline.
A brown dog, muzzle gone gray—surely from a life well lived—tries to climb three steps but falters. Her legs give out, and she twists and falls. A Rottweiler limps around a kitchen. A golden retriever pants in a vet’s office, then he’s placed on a table, wrapped in medical tubes. “Bye, buddy,” a voice says off camera. Nearby, a hand picks up a syringe.
This is Dead-Dog TikTok. It is an algorithmic loop of pet death: of sick and senior dogs living their last day on Earth, of final hours spent clinging to one another in the veterinarian’s office, of the brutal grief that follows in the aftermath. One related trend invites owners to share the moment they knew it was time—time unspecified, but clear: Share the moment you decided to euthanize your dog.
The result is wrenching. A dog is always dying, and someone is always hurting. Likes and sympathetic comments amass. The video goes viral. Engage with one—or even just watch it to completion—and you may be served another, and another. Suddenly, you’re stuck in a corner of TikTok you’d rather not see.
“TikTok has to figure out a way to separate dog content from ‘my dog died’ content,” one user observes in a video from February. He says he can’t stand watching the latter, and his comment section is filled with people agreeing. “The amount of dogs I’ve never met that I’ve cried over is unreal,” one writes.
Dead-Dog TikTok gets at a core tension of the platform writ large. TikTok collapses social media and entertainment, and gives an outsize power to its “For You”–feed algorithm: The user has limited control over what shows up on their feed. Unlike, say, on Reddit, where you might enter a rabbit hole by choice (maybe because you’ve subscribed to the True Crime forum), TikTok’s algorithm might throw you down one based on metrics that may not signal your actual interest.
And in the case of Dead-Dog TikTok, the algorithm can’t know what it means to plop a stranger’s pet loss next to a teen bopping to the latest viral trend or a snippet from late-night television. It can’t recognize that a user’s intention behind posting their dog’s last moments—for catharsis, for validation, to find other people who have felt that same loss—may not be a match for many viewers on the other side who are just trying to pass some time. “We often ascribe all sorts of intentions to the algorithm, like, Oh, it knows,” Nick Seaver, an anthropology professor at Tufts University who studies algorithms, told me. “But it really doesn’t.”
The tension is unresolvable, which is possibly why TikTok rolled out a feature last week allowing users to “start fresh” with a new feed. TikTok, for its part, sees the solution as diversifying the content. “In addition, we work to carefully apply limits to some content that doesn’t violate our policies, but may impact the viewing experience if viewed repeatedly, particularly when it comes to content with themes of sadness, extreme exercise or dieting, or that’s sexually suggestive,” the company wrote in a blog post.
Whatever equation powers TikTok’s For You feed appears to have picked up that videos about dead dogs engage users. But it doesn’t seem to know when to stop serving it, and it tends to go too far, perhaps even by design. “When it finds something that works, it will go and try to push that—both at the individual level and the overall ecosystem level—pretty far,” Kevin Munger, a political scientist at Penn State who has studied the TikTok algorithm, explained to me. “It’s not going to stop at the right level.” To use a positive analogy, it’s as if the algorithm has figured out that you like cake, and so it’s serving you cake for breakfast, lunch, and dinner.
An algorithmic reset may not be able to totally solve this problem—in theory, the app will relearn what you like and serve videos accordingly. Some of the researchers I spoke with said that they very intentionally—even aggressively—signal to the platform what they do and don’t like. When they see a video of a type they don’t want more of, they take action: swiping away quickly, seeking out positive videos, reporting the upsetting content, even closing the app altogether. Other options include blocking specific users or hashtags, or pressing the “not interested” button.
As Robyn Caplan of the Data & Society Research Institute pointed out, an algorithm “can’t necessarily tell the difference between something that is making you cry and something that is making you laugh.” She told me she once asked a friend for funny videos to help “cheer up” her feed.
Grief is a nuanced human experience. “There’s not an obvious context in which you might want to watch videos about pet grief,” Seaver said. “And so it totally makes sense that these systems do these kind of clunky moves, because I don’t think there’s a non-clunky way to do it.” At its best, Dead-Dog TikTok may offer a support community to people suffering and normalize their pain.
Take Blaine Weeks. Weeks thought she had more time with her dog Indica—a few weeks or months, maybe. He was old, and his body seemed to be failing. Then one day, he didn’t want to get up. “I felt like I just didn’t have enough,” she told me. “I didn’t have enough pictures, I didn’t have enough videos, and I was distraught about that.” Weeks decided to record Indica’s last day, worried that otherwise she might block it out entirely with grief.
In the video montage of that day, which Weeks posted to TikTok, she loads Indica into her truck, and they get McDonald’s burgers as a final treat. Weeks tells him that she loves him as he licks tears from her face. Later, on the floor of the vet office, Indica perks up enough to eat a few fries, before resting his head in Weeks’s lap. It ends there.
The post has been viewed 13 million times and climbing. “Randomly last night that video started going crazy again and got, like, another 400,000 views,” she told me when we talked earlier this month. Weeks said that she’d had to turn off her phone for a bit because of negative comments on the video (detractors questioned Weeks’s decision to euthanize) but that overall she’d found comfort from the experience. The video, she said, connected her with more than a dozen people whom she can talk with about her grief. “We kind of check on each other back and forth, saying, ‘Hey, are you doing okay today?’ ‘Yes, I’m doing okay. How are you?’” A stranger made a painting of Indica and sent it to her.
Stefanie Renee Salyers’s TikTok saying goodbye to Princess, her Shih Tzu, has been viewed 28 million times and has nearly 90,000 comments. Salyers got so many messages after posting that she created a Google Form for other people to share their dog-grief stories, offering to read them privately or—with their permission—create TikToks about their lost pets. “I felt, I guess, glad that, even though my video is of a very sad event, that there were people who saw it and felt like, I’m not alone in feeling this grief. And I’m not crazy for feeling like I lost a family member,” she told me.
Crystal Abidin, the founder of the TikTok Cultures Research Network—a group that connects scholars doing qualitative research about TikTok—and a professor at Curtin University, in Perth, Australia, has been studying the comment sections on TikTok grief posts at large. She has found “a really beautiful ethos of care work happening” there: people comforting one another, resource sharing, and more.
Videos like Saylers’s and Weeks’s may inspire others to post their own pet-loss stories. Abidin believes that the pandemic really mainstreamed videos about grief and death on the platform—videos from individuals, videos from health-care professionals. “There is a whole collision of these histories and people of different standpoints and expertise, all on GriefTok,” she told me. “It’s not bad; it’s not good. It’s just that you cannot choose what you want on your feed. And that can be arresting for a viewer.”
Dead-Dog TikTok may be a genuinely helpful space for some, and an upsetting one for others. The platform can’t perfectly sort who’s who. “But if we think about your personal ethos, principles, and morality, do we really want platforms to be the arbiter of what we should and shouldn’t see?” Abidin asked. Maybe TikTok could be smarter about not circulating distressing content, but should it really decide who grieves online and how?
Grief is messy and complicated and hits different people in different ways. So it is only natural that its manifestations online would likewise be messy and complicated. To grieve is to be human—one thing that algorithms, no matter how eerily attuned to our interests and desires, never can be.
The Insidexpress is now on Telegram and Google News. Join us on Telegram and Google News, and stay updated.