Social media algorithms have become highly efficient at detecting emotional signals and delivering content that provokes reaction, including material that may be harmful to young users.
According to Arno Solin, professor at Aalto University, the targeting capabilities of today’s algorithms operate at a level that borders on invisible. “The logic is blind but effective. The algorithm detects what catches a user’s attention and pushes more of it, regardless of the content’s nature,” he said.
The issue resurfaced this week as Finnish police investigated possible links between recent deaths in Helsinki’s Jätkäsaari district and a viral trend on TikTok. One of the deceased teenagers had referenced dark and introspective content on her social media profile.
While Solin declined to comment on specific cases, he emphasised how platforms are designed to identify and serve content that generates engagement, not well-being.
“Content that elicits a reaction is not necessarily content that makes you happy,” he said.
Algorithms can now interpret even the most subtle behavioural signals. Watching a video for a fraction longer than others may be enough to trigger a feedback loop where similar content is prioritised.
“It doesn’t require you to like, follow or search for anything,” said Solin. “That’s often unclear to users.”
Users technically have options to influence the content they see, such as hiding specific creators or flagging content they don’t want, but these tools require active effort. In contrast, algorithmic targeting happens passively and without notification.
“This imbalance makes it difficult for users, especially teenagers, to control what they’re exposed to,” Solin said.
Neuropsychologist Tuomo Aro also sees targeted content as a mental health risk for young people. “If a teen is already feeling low, the algorithm may pick up on this and begin pushing content related to depression or similar themes,” he said.
According to Aro, content served to children under secondary school age should fall under parental responsibility. But he also encouraged older teens to recognise which content worsens their mood.
“Being able to identify that is a useful skill,” he said.
Recent studies and incidents have prompted renewed scrutiny of algorithm-driven platforms. Health professionals across Europe have flagged the psychological effects of algorithmic content cycles, particularly among adolescents.
In Finland, public health authorities have raised concerns over young users being drawn into online subcultures or viral trends that promote self-harm, fatalism or conspiratorial thinking.
Experts stress that algorithmic logic does not differentiate between entertainment and emotional harm. It simply delivers what holds attention longest.
“You don’t need to search for something dark,” said Solin. “It might find you first.”
HT
