Evidence for Social Media Legislation Strategies

Abstract

Importance

Social media use has grown rapidly, with particularly high adoption among adolescents. These patterns carry both risks, such as problematic use, and potential benefits, such as social connection, creating challenges for policymakers developing regulatory approaches.

Objective

To synthesize existing evidence and policy approaches related to legislative efforts addressing adolescent social media use and to assess the potential risks, benefits, and evidence gaps associated with commonly proposed or implemented policy strategies.

Evidence Review

Evidence was identified through database searches from PubMed and Web of Science focused on policies and legislative proposals from August 1, 2003, to March 22, 2026, targeting adolescent social media use. Policy categories included warning labels, age restrictions and assurance (eg, verification), parental consent and monitoring, school phone policies, platform design regulations, and legislation addressing deepfakes and nonconsensual intimate images.

Findings

The search yielded 672 articles, of which 39 met inclusion criteria. Most publications addressed parental consent and monitoring (n = 12), deepfakes and nonconsensual intimate images (n = 8), age restrictions and assurance (n = 7), and design bills (n = 6), with fewer articles examining school phone policies (n = 4) and warning labels (n = 2). Warning labels may increase awareness of potential risks, although data largely consisted of expert opinion. Age assurance measures were commonly circumvented, and proposed biometric or artificial intelligence (AI)–based verification methods had unresolved challenges. Parental monitoring and consent were associated with lower problematic social media use and positive adolescent outcomes in survey-based publications. School phone restriction policies were widely adopted, but evidence regarding academic, social, or mental health effects was limited. Design-focused legislation, such as the Kids Online Safety Act and Age-Appropriate Design Bills, emphasized platform accountability, transparency, and safeguards for minors. Policies addressing AI-generated sexual deepfakes and nonconsensual intimate images aimed to reduce exposure to psychologically harmful material, although evaluation of their effectiveness was limited.

Conclusions and Relevance

In this review of legislative approaches to adolescent social media, observational and expert-opinion publications provided preliminary insights into potential benefits and risks, but systematic evidence evaluating the effectiveness and unintended consequences was limited. Rapid research, including timely evaluation of newly enacted legislation, is needed to assess the effects of different policy approaches and support the iterative, collaborative development of theory-driven, child-centered policies.

Introduction

Social media use has become nearly universal among adolescents, with 95% of those aged 13 to 17 years reporting use.¹ Social media use offers potential psychosocial benefits, including social connection, but is also associated with potential risks, including adverse mental health outcomes.² A key concern of social media use is a phenomenon described as “human fracking” or “attention fracking,” referring to the commodification of attention by technology companies and content creators for profit,³ which can lead to addiction and negative mental, physical, and social challenges.⁴

Approaches to healthy social media use can be broadly characterized as individual-level or system-level approaches. Individual-level approaches rely on individual behavior change, such as parents restricting or banning children’s social media use. In contrast, system-level approaches operate independently of individual decision-making and include measures such as platform-level age assurance (eg, verification) or default access restrictions. Because they do not require sustained user effort, system-level approaches may be more effective at a population level and can help support parents attempting to establish controls amid the many challenges of enforcing parental restrictions.

Policymakers face the critical task of addressing the multifaceted challenges surrounding social media use by implementing protections that safeguard adolescents from harm while preserving the documented benefits, particularly for marginalized youth who may rely on digital platforms for connection, identity exploration, and access to information.⁵ An overview of the current US social media policy landscape is presented in Table 1. Although empirical evidence can inform legislative decision-making, the evidence relevant to proposed and enacted legislation addressing adolescent social media use has not been synthesized in a policy-focused context. This Special Communication synthesizes empirical studies, policy analyses, and other publications from database searches, as well as relevant historical and global context, to examine proposed approaches to regulating adolescent social media use.

Table 1.

US Policies Addressing Adolescent Social Media Use

Policy category	Examples of existing or proposed legislation	Status of bills	Key provisions
Warning labels	State-level bills requiring the display of warning labels before using social media (eg, Minnesota, California)	Enacted in some states, such as Minnesota and California	Require social media platforms to display warnings on opening the app about mental health harms of social media use
Age restrictions and assurance	Federal proposals requiring age verification; state-level requirements to verify ages before creating social media account (eg, Arkansas, Utah, Georgia)	Enacted in several states, but some have been struck down over the First Amendment	Require platforms to verify user age before account creation; set nighttime curfews for minors’ use of social media; aim to reduce exposure to harmful or addictive content
Parental consent and monitoring	State-level social media parental consent laws (eg, Utah, Arkansas, Texas)	Enacted in states, including Arkansas, Utah, and Texas, but some have been struck down over the First Amendment	Require parental consent for minors to download social media apps or create social media accounts; require platforms to allow parents to monitor children’s social media use
School phone policies	State-level laws restricting phone use in schools; district-level policies and bans on smart phones	Phone restriction laws enacted in many states; many school districts have also adopted local policies	Restrict phone use during classroom instruction and school day; require districts to adopt a policy regarding the storage of cell phones
Design bills	Federal Kids Online Safety Act; State Age-Appropriate Design Codes (eg, California, Maryland, Vermont)	Enacted in some states but face legal challenges over First Amendment claims	Require the highest privacy settings by default; require platform risk assessments and safety-by-design standards; require “safe” recommendation systems for minors; allow opt-out of algorithmic feeds
Bills addressing deepfakes and nonconsensual intimate images	Federal TAKE IT DOWN Act; state-level laws outlawing explicit deepfakes and providing pathways to take down intimate images (eg, California, Idaho, Texas)	TAKE IT DOWN enacted on the federal level and state bills enacted	Provide right of action for users to report and take down deepfakes and nonconsensual intimate images; criminalizes nonconsensual explicit deepfakes

Methods

We conducted a search to identify legislation and policies addressing US and global adolescent social media use across 6 themes: warning labels, age restrictions and assurance, parental consent and monitoring, school phone policies, design bills, and bills addressing deepfakes and nonconsensual intimate images. Searches were performed in PubMed and Web of Science for articles published between August 1, 2003 (launch of the social networking service Myspace; Viant Technology LLC), and March 22, 2026. Keywords used to combine concepts of adolescents, social media, and policy included child, youth, minor, adolescent, social media, digital media, online platform, social network, social platform, legislation, law, regulation, mandate, bill, age verification, parental consent, parental monitoring, phone policy, phone ban, harmful content, warning label, and others and were joined using Boolean operators.

We excluded publications not focused on adolescents, nonpolicy editorials, and sources lacking legislative relevance. The review included publications with various study designs, such as empirical studies, cross-sectional and longitudinal surveys, and narrative and scoping reviews. References from included documents were hand-searched for additional relevant sources. Titles and abstracts were screened by 2 reviewers (S.N. and S.Z.), with full-text review to confirm eligibility. A third reviewer (O.H.H.) assessed the sources and checked for potential discrepancies. Any discrepancies were resolved through discussion among coauthors. Data extracted included authors and year of publication, source type, methods or evidence basis, and key findings (Table 2). Additional details are provided in eMethods and eFigures 1 to 6 in Supplement 1. The University of California, San Francisco Institutional Review Board determined this study to be non–human participant research and therefore exempt from the need for approval and informed consent.

Table 2.

Characteristics and Key Findings of Publications on Adolescent Social Media Policy Interventions

Source	Source type	Methods or evidence basis	Key findings
Warning labels
Abbasi and Hswen,6 2024	News or commentary	Expert opinion and narrative evidence	Warning labels may increase awareness and prompt behavioral changes.
Tanne,7 2024	News or commentary	Expert opinion and narrative evidence	Warning labels may increase awareness and prompt behavioral changes.
Age restrictions and assurance
Asher et al,8 2024	Commentary or perspective	Expert opinion and narrative evidence	Adolescents are susceptible to risky social media challenges due to neurodevelopmental and psychosocial factors, which can lead to serious harm, legal consequences, and possibly psychiatric involvement or criminal prosecution.
Costello et al,9 2016	Commentary or perspective	Expert opinion and narrative evidence	Adolescents younger than 18 years can currently enter agreements with social media companies; proposes additional protections for those aged 13 to 17 years.
Köhler-Dauner et al,10 2025	Commentary or perspective	Expert opinion and narrative evidence	Integrated digital child protection approach, including privacy-preserving age verification, age-appropriate platform design, age-tiered access, and media education.
Moreno and Radesky,11 2024	Commentary or perspective	Expert opinion and narrative evidence	Adolescent social media legislation should prioritize platform design, including default privacy settings, user control, limits on data collection, and safeguards against harmful content, rather than strict age restrictions or parental gatekeeping.
Prakash,12 2025	Commentary or perspective	Expert opinion and narrative evidence	Social media use among Indian adolescents presents significant mental health risks; while age restrictions could mitigate these harms, effective policy requires a balanced approach integrating education and regulation of platforms.
Schneble et al,13 2021	Original research	Thematic analysis of terms and conditions of 22 mobile platforms	Social media platforms inadequately obtain informed consent from children; recommends age-appropriate verification and parental involvement.
Stevens et al,14 2026	Commentary or perspective	Expert opinion and narrative evidence	Australia’s social media ban for users younger than 16 years is unlikely to be effective alone and requires rigorous evaluation complemented by additional regulatory measures.
Parental consent and monitoring
Al-Shoaibi et al,15 2024	Original research	Survey of adolescents and parents (n = 10757)	Parental monitoring linked to lower adolescent screen time; family conflict linked to higher screen time; associations with video games, video-sharing platforms, and television and movies.
Beyens et al,16 2022	Review	Narrative review	Active, autonomy-supportive parental mediation linked to reduced problematic use, depression, anxiety, and cyberbullying effects; restrictive mediation linked to worse outcomes.
Cohen et al,17 2023	Original research	Survey of parents (n = 1613)	US parental monitoring varies with child age and bullying concern—higher among younger, bullying-concerned parents.
Hernandez et al,18 2024	Original research	Mixed methods analysis: survey and qualitative interviews	Restrictive parental monitoring of early adolescents’ digital media use was linked to higher problematic internet use, whereas active and deference monitoring were associated with positive family dynamics.
Lahti et al,19 2021	Original research	Survey of adolescents (n = 3408)	Finnish adolescents’ internet use was categorized into 5 distinct profiles, each linked to demographic, social, and health characteristics, including gender, parental monitoring, peer support, and well-being.
Morawska et al,20 2026	Review	Systematic review	Parental influence varies: modeling and mediation were beneficial; monitoring and limits were mixed.
Nagata et al,21 2025	Original research	Survey of adolescents and parents (n = 10048)	Parental monitoring associated with lower problematic social media use and lower total screen time.
Prendergast and Dyer,22 2026	Original research	Systematic review	Media coverage of Australia’s social media age restrictions portrays youth as vulnerable and passive; underscores need for policy approaches that center young people’s agency and broader social contexts in promoting well-being.
Schwarz-Torres et al,23 2025	Review	Narrative review	Adolescent parental monitoring strategies; gaps in prioritization (eg, when to decrease monitoring, how to monitor using technology).
Vigil et al,24 2025	Original research	Survey of parents (n = 1359)	Parents generally supported social media legislation requiring parental consent, full account access, and data privacy policies, and believed they positively impact adolescents’ well-being.
Walch and Sabey,25 2020	Review	Narrative review	Proactive parent monitoring and discussion of social media use can lead to positive emotional outcomes.
Wallace,26 2021	Original research	Survey of parents (n = 389)	US Pennsylvania study of maternal vs paternal monitoring differences found mothers monitored adolescent social media use more frequently than fathers; both parents had greater monitoring of daughters than sons.
School phone policies
Cantor et al,27 2025	Original research (short report)	Survey of nationally representative school principals (n = 985)	A total of 96.7% of schools had a cell phone policy; middle and high schools had less strict policies than elementary schools.
Goodyear et al,28 2025	Original research	Survey of adolescents (n = 1227)	English secondary schools that implemented phone restrictions had reduced in-school use, but no change in overall use or well-being.
King et al,29 2024	Original research	Survey of adolescents (n = 1282)	South Australia school phone ban had no short-term effects on problematic use, academic engagement, or school belonging; minor bullying differences.
Tandon et al,30 2020	Original research (short report)	Survey of middle and high school principals	More than 90% of principals supported phone restrictions; more than 80% believed phone use during school negatively affects social development and academics.
Design bills
Alemi et al,31 2023	Letter	Expert opinion and narrative evidence	Partial support for Kids Online Safety Act; emphasizes protecting youth from harmful content and research; notes need to avoid censoring critical content for LGBTQ+ youth.
Chhabra et al,32 2025	Original research	Scoping review and content analysis of 70 reports of policy and platform recommendations	Recommendations emphasize accountability, transparency, and limiting youth access to social media.
Costello et al,33 2023	Original research or interdisciplinary study	Mixed-methods analysis: empirical algorithm testing, economic analysis, literature review, legal review	Social media algorithms expose youth to harmful content; economic incentives drive amplification of this content; policy may mitigate harms.
Fatt and Fardouly,34 2025	Commentary or perspective	Debate and commentary	Australia social media restrictions on users younger than 16 years; debate on effectiveness; targeted design regulation preferred over age bans for adolescent mental health.
Livingstone and Sylwander,35 2026	Commentary or perspective	Expert opinion and narrative evidence	Mixed evidence on social media harms; limited support for fixed age limits; flexible, context-dependent approaches supported by developmental science and child rights.
Welsey-Smith and Fleming,36 2025	Commentary or perspective	Debate and commentary	Adolescent social media use linked to consistent harms (eg, distress, disordered eating); supports age protections and design regulation.
Bills addressing deepfakes and nonconsensual intimate images
Chawki,37 2024	Original research	Computer science legal and policy analysis	US policies: Deepfake Report Act (2019), National Institute of Standards and Technology standards, platform flagging or removal of manipulated media; need stronger regulation and state action.
Farid,38 2025	Commentary or perspective	Expert opinion and narrative evidence	US laws criminalize nonconsensual intimate images, including deepfakes; TAKE IT DOWN Act requires 48-hour removal; calls for coordinated laws and stronger enforcement.
Meneses,39 2024	Original research	Legal and policy analysis	Defines key characteristics of deepfakes; Louisiana SB 175 criminalizes deepfakes involving minors and defines digital image and likeness rights.
Najdowski,40 2017	Original research	Legal and policy analysis	US laws make prosecution of nonconsensual pornography challenging; legal reform remains insufficient.
Recupero,41 2016	Commentary or perspective	Expert opinion	Forensic psychiatrists need to understand online harms; weak enforcement due to lack of resources, expertise, or prioritization.
Russell,42 2024	Commentary or perspective	Debate and commentary	Evidence-based content moderation needed to protect youth mental health; algorithmic recommendations frequently expose adolescents self-harm and suicide content.
Thimm-Kaiser and Keys,43 2025	Original research	Legal and policy analysis	Limited evidence of social media harm to adolescent mental health and development; higher risk in vulnerable youth; gap between rapid policy and evidence.
Zhang et al,44 2024	Commentary or perspective	Debate and commentary	Social media moderation of mental health content are often ineffective; may isolate at-risk youth and limit support; need evidence-based policies with mental health input.

Abbreviation: LGBTQ+, lesbian, gay, bisexual, transgender, queer, and other sexual and gender minority.

Results

Our search yielded 672 articles, of which 39 met inclusion criteria following title and abstract screening.^6–44 The largest number of included articles addressed parental consent and monitoring (n = 12),^{15 –26} deepfakes and nonconsensual intimate images (n = 8),^{37 –44} age restrictions and assurance (n = 7),^{8 –14} and design bills (n = 6),^{31 –36} with fewer articles examining school phone policies (n = 4)^{27 –30} and warning labels (n = 2).^{6 ,7} Characteristics and key findings of included publications are summarized in Table 2. Specifically, expert-opinion studies suggested warning labels may increase awareness and behavior change, while age assurance, digital literacy, and platform safety measures would support safer use. Survey studies found that parental modeling and mediation were associated with improved social media use. Table 3 and the eTable in Supplement 1 present current social media design policies aimed at reducing risk among adolescents. A detailed review of these 6 themes of adolescent social media use follows.

Table 3.

US and Global Social Media Design Policies Aimed at Reducing Risk Among Adolescents

Design regulation	Examples of existing or proposed legislation	Key provisions
Restrictions on infinite scroll and personalized feeds	New York Stop Addictive Feeds Exploitation for Kids Act; Utah Minor Protection in Social Media Act	Restricts infinite scroll or requires platforms to display reverse-chronological feeds, instead of content based on recommendation algorithms
Notifications for minors	California Protecting Our Kids from Social Media Addiction Act	Bans notifications for minor users typically at night and/or during the school day
Privacy-by-default settings for minors	Maryland Age-Appropriate Design Code; Utah Minor Protection in Social Media Act	Requires social media platforms to implement the strongest data privacy features for minors by default
“Duty of care” standards	Kids Online Safety Act; Maryland’s Age-Appropriate Design Code; Vermont Age-Appropriate Design Code	Requires social media platforms to design platforms to avoid harm to minors
Dark patterns	Nebraska’s Age-Appropriate Design Code; Vermont’s Age-Appropriate Design Code	Restrict the use of “dark patterns,” user interface choices that lead users to make choices unfavorable to them, in minors
Age-appropriate design and privacy	United Kingdom’s Age-Appropriate Design Code	Highest privacy settings by default for minors; minimize unnecessary data collection; verifiable parental consent for children younger than 13 years
Data collection and platform design restrictions	European Union’s Digital Services Act	Limits on data collection; transparency obligations; platform safety requirements
Data use and access	United Kingdom’s Data (Use and Access) Bill	Restrictions on how platforms collect and use data; adherence obligations for technology companies
Online services regulation	United Kingdom Online Services Act	Requirements to manage harmful content; safety and design standards for online platforms

Warning Labels

Warning labels have historically been used to communicate potential health risks to consumers. Such labels have been implemented for products such as tobacco, alcohol, and cannabis, for which health risks are well characterized. In 2024, then–Surgeon General Vivek Murthy recommended implementing warning labels on social media platforms, highlighting growing national concern regarding adolescent mental health.^6,7,45 In contrast to regulated substances, social media is associated with a broader range of potential risks and benefits,^46–49 presenting challenges for the development of labeling strategies that accurately reflect this complexity.

Warning labels have been most extensively developed in tobacco regulation, where evidence demonstrates their effectiveness in communicating health risks and lowering use. Research suggests that tobacco warning labels are associated with changes in risk perception and smoking behavior.^50–52 Unlike tobacco, the trajectory of alcohol warning labels has been limited, with the 1989 Alcoholic Beverage Labeling Act’s text-based warnings now being criticized as outdated and ineffective.^52–54 Collectively, warning labels may influence risk perceptions and behaviors, although evidence for alcohol is limited, and effects vary by content, design, and regulatory context.

These considerations are particularly relevant to social media, where risks and benefits are more heterogeneous than for alcohol or tobacco products. Unlike prior warning label research that has largely focused on individual users, social media warning labels may influence adolescents, parents, and caregivers simultaneously. Potential effectiveness may also depend on platform implementation and adherence, underscoring the importance of regulatory oversight.

Social media use among adolescents reflects a dynamic balance between potential benefits and harms, influenced by patterns of access and use. Although higher use has been associated with anxiety, depression, and exposure to harmful content, it also supports social connection, identity exploration, and mental health, particularly for marginalized adolescents.^6,7,46 Warning labels may influence behaviors that reduce exposure to harmful content, but concerns remain that overly cautionary or fear-based messaging may affect engagement with beneficial uses. Accordingly, existing evidence suggests that the design and framing of warning labels warrant careful consideration to minimize unintended consequences while addressing potential harms. Rather than solely listing the potential harms of social media use, warning labels should also provide resources and strategies to reduce risky use; in this way, warning labels can also serve as a resource hub to support adolescents’ well-being.

Age Restrictions and Assurance

In the US, most social media platforms require users to be at least 13 years of age to create an account, a threshold often termed the age of digital consent.⁵⁵ Policy discussions have considered whether more legal protections are needed for adolescents aged 13 to 17 years, who are not recognized by the law as adults and may differ from adults in developmental and social characteristics that may influence heightened reward sensitivity, increased sensation seeking, and peer influence.^8,9,13 Platforms such as Instagram (Meta Platforms), Facebook (Meta Platforms), TikTok (ByteDance), X/Twitter (X Corp), Snapchat (Snap Inc), and other social media sites have age restriction rules.⁵⁶ These rules and regulations stem from the 1998 Children’s Online Privacy Protection Act, which restricts social media companies from collecting personal information from children younger than 13 years without parental consent.⁵⁷ Although the Children’s Online Privacy Protection Act does not specifically prohibit children younger than 13 years from creating social media accounts, the design of social media requires collection and processing of user data to an extent that would be prohibited under this law. However, evidence suggests that these age minimums are routinely circumvented, as 64% of children aged 11 and 12 years report social media use, and underage users report having 3.4 social media accounts on average.⁵⁶ These findings suggest limitations in the effectiveness of self-declared age-gating on social media.

Although age restriction rules are in place, accurately verifying user ages presents challenges. While most platforms comply with mandated age restrictions, age assurance systems often rely on reactive estimation of a teen’s age rather than proactive verification.¹³ Previous literature indicates that self-declared age-gating can be bypassed through disposable email addresses or browser refreshes.⁵⁸ Most social media platforms now use artificial intelligence (AI)–based age estimation based on the posts and information shared by users to reactively identify underage users. Other approaches being tested include identity-based verification, such as requiring government-issued identification or biometric-based age estimation.⁵⁹

Biometrics identify individuals using unique physical and behavioral characteristics and are already used in some clinical settings through integration with information systems technology.⁶⁰ Commercial products using facial biometrics have shown variable performance, which may reduce the feasibility for social media applications.⁶¹ Additionally, AI is an emerging technology being used to regulate online age assurance. Pretrained AI models, such as VGG16 (Visual Geometry Group; University of Oxford) and ResNet50 (Microsoft Research), build upon larger datasets and can estimate age more efficiently, with most current models using regression models instead of classification models to improve accuracy.⁶² However, variation in aging related to lifestyle, health status, and ethnicity remains difficult for AI to capture, resulting in misclassification risks.⁶²

Other emerging technologies for age verification include AI-based age estimation, optical character recognition for government identification, and digital parental consent workflows. Layered approaches in which nonintrusive AI checks are followed by documentation only when necessary may balance adherence, usability, and privacy. Interoperable methods, such as digital identity wallets, may allow repeated verification without sharing data multiple times.

Age assurance may support adherence to minimum age policies, but biometric identification using facial recognition may raise privacy and safety concerns. In particular, members of sexual and gender minority groups whose appearance may not match their government-issued documentation and/or sex assigned at birth may be misidentified, potentially limiting access.

Parental Consent and Monitoring

Parental monitoring encompasses parental supervision, child disclosure, active monitoring efforts, and parental knowledge.⁶³ Both child characteristics and parental concerns regarding peer victimization can shape monitoring approaches.¹⁷ Increased parental experience with social media platforms may also be associated with increased surveillance of their child’s online activity.⁶⁴ Monitoring practices have been associated with patterns of social media use that may support positive developmental outcomes²⁵ and with lower levels of problematic social media use.²¹ Conversely, low levels of parental supervision, particularly in conjunction with elevated parent-adolescent conflict, may be associated with a greater risk for social media addiction later in adulthood.⁶⁵

Studies have shown that 81% of US adults, compared with 46% of teens, support policies requiring parental consent for minors to create social media accounts,⁶⁶ and parents generally believed parental consent and monitoring would have a positive impact on adolescents’ well-being.²⁴ In 2025, Utah became the first state to implement an age verification law through the App Store Accountability Act, which mandates parental consent before minors can download apps.⁶⁷

The effects of legislation requiring parental consent and allowing parents to monitor or access their children’s social media accounts may vary across adolescents. Those living in unsupportive home environments may rely on social media as a safe space for affirmation and social belonging.⁶⁸ Policy decisions should therefore balance efforts to reduce online risks with preserving access to supportive online communities. In addition, legislation allowing parents to access adolescents’ social media accounts may be insufficient on its own and should be complemented by platform-level safeguards.

School Phone Policies

Restrictions on use of phones or personal electronic devices in schools are becoming increasingly common across the US, with at least 38 states and Washington, DC, enacting restrictions.⁶⁹ While some states, such as Florida, have implemented strict statewide bans prohibiting phone use during instructional time, others, such as California, allow individual school districts to develop their own policies. Beyond state mandates, many local school districts have implemented phone restrictions independently.

A primary objective of phone bans is to improve academic performance; however, evidence is mixed. Some studies report modest increases in grade point average and student concentration following bans on personal devices,^70,71 while other research has found no significant change in overall academic outcomes.⁷²

School phone bans have also shown inconsistent associations with phone use and mental health. For instance, Australian schools reported reductions in problematic phone use after implementing school-based phone bans,²⁹ while an English study of adolescents aged 12 to 15 years found no significant differences in mental well-being between schools with restrictive phone bans and those without.⁷³ Moreover, some surveys indicate that adolescents in schools with phone bans report lower well-being than those in schools without restrictions.⁷⁴ Phone restrictions have been associated with academic benefits, including higher science and math scores.⁷⁵ These findings underscore the need for further research to clarify relationships between school phone bans, academic outcomes, and student mental health.

Schools also need clear rules and procedures to implement these policies effectively. For instance, policies may require phones to be stored in lockers, magnetic pouches, or left at home. Schools must also address student adherence with the rules, as monitoring and enforcing bans when a sizable proportion of students continue to use phones in class can reduce instructional time. Exceptions are also necessary; for example, students who use their phones to operate medical devices such as insulin pumps must be permitted access during school. Educators have reported perceived benefits of phone policies, including reduced classroom distractions and improved engagement.⁷⁶ These policies only regulate in-school use, highlighting the potential role of complementary strategies to manage screen use outside of school.

Design Bills

The rapid expansion of digital technologies has been associated with increased psychological distress, exposure to abuse, and exposure to violence among children and adolescents,⁷⁷ prompting interest in social media platform design. Many design features embedded within social media platforms are intended to maximize engagement and may contribute to harmful use patterns and abuse. For example, the infinite scroll feature may promote “doomscrolling,” a pattern of compulsive consumption of large volumes of often negative or distressing content,⁷⁸ which has been associated with social media addiction, fear of missing out, and other maladaptive patterns of use.⁷⁹ Algorithmic amplification—the use of algorithms to manipulate and influence exposures to specific types of content on social media platforms—represents another design feature with potential adverse effects.⁸⁰ Exposure to algorithmically amplified content has been associated with elevated rates of mental health concerns in adolescents, including body image concerns, eating disorders, and suicidality.³³

The Kids Online Safety Act (KOSA) was first introduced in the US in 2022. This bill seeks to safeguard minors against negative outcomes, including depression, problematic social media use, online harassment, and sexual exploitation, among others.⁸¹ Under KOSA, platforms would have a “duty of care” to mitigate harm for children.⁸¹ The bill would require platforms to implement default safety settings for children and restrict highly personalized recommendation systems that rely on children’s data to deliver algorithmically curated, infinite content feeds. Under the proposed legislation, minors would be able to opt out of personalized feeds in favor of chronological content.⁸¹ It would also mandate independent, third-party audits to evaluate platform adherence.⁸¹

At the state level, similar legislative efforts have emerged through age-appropriate design bills in states such as Nebraska, Vermont, and California.⁸² These laws include a variety of design-focused provisions, including limits on infinite scrolling, restrictions on push notifications for minors, prohibitions on dark patterns, and strengthened privacy standards. Although many design bills were signed into law, they have faced constitutional challenges. Federal courts have halted certain measures of California’s Age-Appropriate Design Code due to First Amendment concerns.

Bills Addressing Deepfakes and Nonconsensual Intimate Images

The advent of AI has expanded the scope of digital harms, enabling the creation of explicit deepfake images. Data indicate that approximately 1 in 17 adolescents has been the subject of an explicit deepfake.⁸³ Additionally, 31% of adolescents report awareness of nonconsensual intimate images being shared within their school, and 15% report hearing about deepfake versions of such images. Most of these images are reportedly shared via social media platforms.⁸⁴

Research on the impact of deepfake intimate images on adolescent mental health is limited. Prior studies have identified negative outcomes, including a higher likelihood of seriously contemplating and attempting suicide,⁸⁵ as well as increased exposure to bullying and harassment.⁸⁶ To address these potential harms, many states have introduced new legislation aimed at deepfakes; currently, 47 states have laws restricting deepfakes, and 45 states have laws prohibiting the creation of AI-generated child sexual abuse material (CSAM).^87,88 Laws such as California’s AB 1831 expand the definition of child pornography to include CSAM.⁸⁹ Because these laws were recently enacted, further research is needed to evaluate their effectiveness in restricting the spread of nonconsensual intimate images.

At the federal level, the TAKE IT DOWN Act strengthens protections by prohibiting the knowing publication of nonconsensual real or deepfake sexual content. The act also provides recourse for victims and requires platforms to remove content within 48 hours of a victim’s request.^38,90 Adherence to mandated removal timelines may be inconsistent, based on prior challenges in enforcing nonconsensual image and CSAM removal.

Discussion

Despite the growing popularity of adolescent social media use policies and practices such as warning labels, age restrictions, school-based phone bans, and parental consent requirements, rigorous research evaluating their effectiveness remains limited. Key priorities for future research are summarized in Table 4 and include improving the measurement of adolescent social media exposure and evaluating how specific policy components affect adolescent social media use and outcomes. Although these evidence gaps remain, policymakers often must act before definitive evidence is available. Policymakers and researchers can adopt a minimal viable evidence model in which emerging evidence of potential harms is used to inform policies designed to be rapidly evaluated and iteratively refined.⁹¹ As evidence accumulates regarding both the prevalence of harms and the effectiveness of regulatory interventions, policies can be revised to better promote adolescent well-being.

Table 4.

Future Research on Social and Digital Media Health Policy

Area	Key need	Recommended methods for future studies
Exposure measurement	Improve measurement of adolescent social media exposure, identify harmful content, and assess platform-specific effects linking exposures to outcomes.	Passive data collection; retrospective measures for cross-sectional studies; repeated, prospective measures for longitudinal studies; platform-specific analyses to identify harmful content.
Harm assessment	Understand how algorithm or social media policy changes affect social media exposure and mental health outcomes in the short and long term at individual and population levels.	Case-control studies with retrospective social media exposure for rare harms; prospective longitudinal studies of social media use trajectories and mental health outcomes for common harms; experimental studies of social media design.
Algorithms and design	Understand how algorithm changes and policy implementation influence social media exposure and outcomes.	Randomized clinical trials of algorithm changes; platform intervention evaluation; population-level analyses using standardized mental health surveys to examine region-wide policy implementation.
Research funding	Government support needed for large-scale research on social media health effects across adolescent developmental stages; include adolescent involvement in study design, interpretation, and safeguard development.	Federal agencies (eg, National Institutes of Health), private foundations, and nonprofits to fund and support research in this area.
Role of industry	Assess industry response to policy, including their influence on policy development and adherence to new regulation.	Study and monitor the industry, their responses to development of regulatory requirements (or attempts) that are implemented and how they shape the types of policies developed.
Policy development and implementation	Study how policymakers design social media policies to reduce harm, including evidence used, intervention type, target groups, policy alignment, resources for implementation, and monitoring fidelity.	Research on policy impacts and combined effects across levels. Create a national repository of resources, model laws, research, and best practices, and provide technical assistance to speed up implementation and scaling of strategies.

Limitations

This study has some limitations. The search strategy did not include conference papers, dissertations, and gray literature, which may have excluded relevant policy analyses and emerging evidence not yet published in peer-reviewed journals. The search was limited to PubMed and Web of Science, potentially omitting relevant publications from other databases. Despite broad search terms, relevant publications may have been missed due to variability in terminology.

Conclusions

In this review of legislation concerning use of social media by adolescents, observational and expert-opinion evidence suggest potential benefits for awareness, monitoring, and risk mitigation. New US state and federal policies, as well as global legislation, aim to mitigate potential social media harms, although evidence regarding these policies’ effectiveness remains limited. Research on policy effectiveness is limited by funding constraints, platform barriers, and legal restrictions on data access. Moreover, data on social media policy implementation have not been systematically synthesized, limiting opportunities to assess impacts such as population reach and intended effects. Evidence gaps highlight the need to assess potential benefits and risks before implementing new social media policies. For example, warning labels may discourage adolescent social media use, but they may also paradoxically drive curiosity and increase social media use. Furthermore, policymakers should note that interventions are often studied in isolation, although multiple interventions may be implemented simultaneously. When multiple interventions occur simultaneously, it may be difficult to discern which is more effective or potentially harmful. Understanding the interactions between interventions and considering youth perspectives may inform more effective policy development.

Key Points.

Question

What legislative and regulatory approaches address adolescent social media use, and what evidence informs their risks and benefits?

Findings

This Special Communication synthesized 39 empirical studies, policy analyses, and historical context to examine warning labels, age restrictions and assurance, parental consent and monitoring, school phone policies, platform design regulations, and policies addressing deepfakes and nonconsensual images. Observational and expert-opinion evidence suggested potential benefits for awareness, monitoring, and risk mitigation, but further research is needed to determine the effectiveness, feasibility, and unintended consequences of particular policy strategies.

Meaning

These findings suggest that current legislative approaches to adolescent social media use outpace available evidence, underscoring the need for evaluation to inform balanced, evidence-based policy development.

Data Sharing Statement:

See Supplement 2.