ABSTRACT
Purpose: Adolescents are using social media now more than ever, especially given the ongoing COVID-19 pandemic. A growing body of research demonstrates longitudinal associations between social media use and poor mental health outcomes among adolescents, including anxiety, suicidality, depression, poor body image, and eating disorders. One driving factor in the role of social media on mental health concerns is social media platforms' use of algorithms, which drive increasingly extreme content to vulnerable youth users over time;however, little is known about what specific policy action could be taken, and on what legal basis, to prevent the harms caused by social media to adolescents. The purpose of this study was to inform the development of practical, evidence-based policy strategies to regulate social media platforms' use of algorithms. Methods: We assembled a team of adolescent health researchers and legal scholars specializing in First Amendment law. Two separate narrative reviews were conducted to inform the legal research and summarize the evidence of the harms of social media on adolescents' (1) neurodevelopment (e.g., cognitive control, reward processing, neuronal activity);and (2) mental health (e.g., eating disorders, depression, cyberbullying). Legal scholars conducted reviews of federal and state legal precedent, with a focus on First Amendment rulings, to investigate avenues for the regulation of social media platforms' use of algorithms. Finally, the adolescent health and legal research teams conducted virtual consultations with stakeholders in business regulation, technology, and federal and state government, including state lawmakers and attorneys general, to inform the findings. Results: The findings of the narrative reviews highlight the significant harm posed to adolescents by social media algorithmic practices. Rigorous studies find that social media use poses negative effects for mental health (e.g., body dissatisfaction, anxiety, depression) among youth who exhibit high use in early adolescence, frequently use image-based platforms, and who are girls. Normal processes of brain development in mid adolescence may heighten vulnerability to exaggerated emotional responses to platform algorithmic practices, such as feeds that are tailored to user behavior and content that becomes increasingly extreme the longer a session continues. These harms provide justification for legal approaches to regulate social media platform practices and protect adolescent mental health. Furthermore, our legal research indicates that states rather than the federal government are more able and likely to adopt innovative and effective legal strategies that could survive constitutional scrutiny (i.e., the First Amendment). One policy innovation that holds the most promise to protect adolescent mental health would be for states to compel social media companies to conduct algorithm risk audits by objective third parties and publicly disclose the results. Algorithm risk audits are a promising avenue to mitigate the harms caused to adolescent mental health resulting from social media platform algorithms. Conclusions: Our legal analysis has generated viable policy recommendations and actions that state lawmakers and attorneys general can take to protect adolescents from the harms of dangerous social media algorithms. Sources of Support: This study was funded by the Becca Schmill Foundation.