TikTok offers belonging to people with mental illnesses, but a University of Minnesota study found that the popular social media app can be a "runaway train" of distressing videos based on how it picks content.
Detailed interviews of 16 TikTok users confirmed the suspicions of U computer scientists: The app's algorithm can draw people together around videos they find interesting, fostering community, but it can do damage if it force-feeds content that increases anxiety.
"Users are along for the algorithmic ride with little ability to affect what the train is doing," they reported. "The train takes them to places they may not themselves want to be about their own past experiences and traumatic events, yet they cannot stop seeing it without a complete cessation of app use."
The U authors said the small study was a first step toward broader research, including solutions that make social media platforms more sensitive to mental health issues and to people who need breaks from their content. The study was promoted Tuesday and will be featured next week at a conference in Germany on human-computer interaction.
TikTok is often minimized as a source of dance and comedy videos, study coauthor Stevie Chancellor said, but it has become a relevant gathering place for mental health support and information. A review of 100 videos with mental health hashtags showed that they had been viewed 1.3 billion times.
People explore TikTok for clinical information, but sometimes they just want practical tips from others dealing with the same kinds of stress or a video of kittens to brighten their days, she added. Informal communities of support grow around videos and content that people like.
"I don't want to throw TikTok out with the bathwater, because it does provide this really useful forum for disseminating mental health information and sharing personal experiences," she said. The U assistant professor specializes in artificial intelligence to identify people in online communities at risk for severe health problems or behaviors.
TikTok's algorithm brings relevant videos from across its platform to people based on their apparent interests, whereas Chancellor said Facebook focuses more on content based on their friends and followers.
The findings about TikTok rang true with clinicians, who discuss with patients how to maximize benefits of social media while minimizing harms.
"When we think about mental health, one of the things people seek out the most is not feeling alone and feeling understood," said Sarah Jerstad, clinical director of psychological services for Children's Minnesota. "This is a way for people to put out their own experiences in small bites about mental health struggles and to get support."
Jerstad advises patients to discuss what they see on TikTok. Sometimes people identify with others on the app who report specific disorders, but then worry that they might have those disorders without telling anyone, she said.
"It's great to bring that to a therapist or a primary care doctor or a caring adult in their life and to be able to say, 'I saw this. I really relate to this. Sometimes I wonder if I struggle with this,'" she said.
Chancellor said the study was small in number because each interview took an hour or two, and common themes quickly emerged. The interviewees ranged in age from 16 to 54 and from working adults with spouses to single teenage students.
All valued TikTok, but they also found that it doesn't change topics, even if they press a button asking to see something else, Chancellor said.
"TikTok is very sticky about content that it believes you like," she said.
Chancellor said she wants to develop AI technology that helps social media platforms identify when people need topic changes. One idea, she said, is to create a "choose your own adventure" option by which the social media app offers entirely new topics or allows people to stick with their current thematic content.
---------
