New Study Could Show How TikTok’s Algorithm Affects Youth Mental Health

A multi-institutional research team led by Georgia Tech has launched a four-year study to investigate how TikTok’s recommendation algorithm influences the mental health and behavior of adolescent users. Funded by a $1.7 million grant from the Huo Family Foundation, the project will analyze data from over 10,000 young people to understand the effects of passive content consumption and 'doomscrolling.' This research is critical for the mental health technology sector as it aims to provide the causal evidence needed to develop design interventions that could prevent the onset of serious mental health conditions in youth.

Led by Professor Munmun De Choudhury of Georgia Tech’s School of Interactive Computing, the study utilizes a $1.7 million grant to audit TikTok’s recommendation engine using data from 10,000 adolescent users in the UK. Collaborating with experts from the University of Cambridge and UCLA, the team will specifically examine watch histories—a shift from traditional research that focuses on active posts—to understand the impact of passive social media consumption. This data-gathering process relies on archives shared with consent under GDPR regulations, bypassing the increasingly restrictive API limitations set by platforms like X (formerly Twitter).

The research aims to characterize the negative exposures young people experience and build computational methods to track how these exposures affect behavior. By using artificial intelligence to simulate realistic video feeds and consumption rabbit holes, the team intends to see how algorithms steer users toward specific content pathways. This methodology follows previous work by UCLA’s Homa Hosseinmardi, who used bots to study YouTube’s recommendation patterns, and will now be applied to TikTok’s influential short-form video model that has been widely adopted by competitors like Instagram and X.

For the mental health technology market, the findings could lead to significant design interventions aimed at minimizing the harmful effects of algorithmic content. De Choudhury noted that because many mental health symptoms first appear during adolescence, identifying unhealthy behavioral patterns early could allow for preventative care and support. The study is part of a broader $17.6 million initiative by the Huo Family Foundation to explore how digital technology shapes brain development and social behavior, addressing a critical lack of causal evidence in the field.