Neurodevelopmental Mechanisms of Adolescent Online Risk: A Multi-Level Perspective on Social Media and Metaverse Harms

Abstract

Background: Adolescents’ engagement with social media and emerging metaverse platforms has become nearly universal, creating environments rich in opportunities for learning, creativity, and social connection. However, these same spaces also enable a range of risky behaviors (RBs) with potential impacts on mental health, safety, and development. Recent research (2022–2025) has documented rising concerns over cyberbullying, online sexual exploitation, self-harm content, problematic use, and new risks specific to immersive VR. Aims: This Perspective uses a narrative synthesis of recent empirical and theoretical literature, including four key articles provided by the author and over 40 additional peer-reviewed and institutional sources, to (i) map the most prevalent and emergent RBs in adolescent social media and metaverse use, (ii) clarify the neurodevelopmental and socio-technical mechanisms that link these behaviors to individual and contextual factors, and (iii) propose a multi-level framework for intervention, policy, and future research aligned with adolescent development. Methods: A narrative synthesis approach was adopted, which is appropriate for integrating heterogeneous study designs and rapidly evolving evidence. The review emphasizes studies published from 2022 to 2025, with a focus on large-scale surveys, longitudinal cohorts, systematic reviews, and scoping reviews relevant to adolescent online risk. Results: Evidence indicates small but consistent associations between high-intensity platform use and internalizing symptoms, with gendered pathways and cultural moderators. Algorithmic amplification contributes to the spread of harmful content, while immersive environments increase the salience and emotional impact of interactions. Certain groups—those with prior trauma, low SES, or marginalized identities—face heightened vulnerability. Conclusions: RBs in digital spaces emerge from the interplay of adolescent neurodevelopment, platform affordances, and socio-cultural context. This Perspective synthesizes recent evidence via narrative review to articulate these mechanisms and to inform an integrated, multi-level framework for harm mitigation that aligns research, platform design, and policy with adolescent developmental needs, while preserving the benefits of digital engagement.

1. Introduction

Adolescence is a developmental stage characterized by profound biological, cognitive, and socio-emotional changes that collectively shape identity formation and future trajectories [1,2]. During this period, adolescents show heightened reward sensitivity and socio-affective responsiveness, combined with still-maturing executive control systems [3,4]. These traits render young people particularly responsive to social evaluation and novelty, both of which are abundant in today’s digital ecosystems [5]. The convergence of developmental predispositions and persuasive digital environments increases susceptibility to both beneficial and harmful online experiences. By 2024, global surveys indicated that over 90% of adolescents in high-income countries use at least one social media platform daily, with significant increases in usage frequency during the COVID-19 pandemic and beyond [6,7]. In parallel, immersive technologies such as virtual reality (VR) and early-stage metaverse platforms have begun to capture the attention of younger cohorts: an estimated 25–30% of U.S. teenagers report monthly use of VR headsets, and ownership is on the rise [8]. This dual engagement with traditional social platforms and immersive environments has created a layered digital landscape where distinct yet interconnected risk vectors operate, spanning content-related harms, contact risks, conduct issues, and monetization-driven exploitation [9]. For the purposes of this Perspective, “risky behaviors” (RBs) are defined as digital activities that increase the likelihood of adverse outcomes for adolescents, including psychological harm (e.g., depression, anxiety), physical harm (e.g., self-injury during challenges), social harm (e.g., reputational damage), and economic harm (e.g., financial exploitation through microtransactions) [10]. This broad definition encompasses the “4Cs” of online risk—content, contact, conduct, and contract—while adding a fifth dimension, embodiment, to capture metaverse-specific phenomena such as immersive harassment and the Proteus effect [11,12]. The theoretical framework guiding this analysis is integrative, drawing from (i) developmental neuroscience, which highlights the temporal gap between heightened socio-emotional drive and mature self-regulation [1,3]; (ii) the affordances approach, which examines how platform features like persistence, scalability, and algorithmic curation shape behavior [13]; and (iii) networked publics theory, which considers how social norms and behaviors emerge and spread within digitally mediated communities [14]. Crucially, very few contributions have explicitly integrated these perspectives to examine how platform-specific affordances—including immersion and embodiment in metaverse environments—interact with adolescent neurodevelopment to shape online risk. This Perspective addresses that gap by linking mechanisms to concrete pathways for mitigation across practice, design, and policy. This Perspective focuses on research published between January 2022 and June 2025, aligning with the journal’s emphasis on situating commentary within the last three years of literature. Evidence is drawn from large-scale cross-sectional surveys, longitudinal cohort studies, systematic and scoping reviews, experimental research, and authoritative institutional reports [7,15,16]. The review also integrates findings from the author’s provided articles, which offer critical insights into cyberbullying dynamics, grooming risk factors, network analysis of harmful discourse, and the psychological interplay between social media and metaverse environments [17,18,19,20]. The scope is deliberately broad to capture the multifaceted nature of RBs across platforms and modalities. Nevertheless, the discussion prioritizes behaviors with robust empirical support and clear policy or design implications. The objectives are threefold: (1) to map the current landscape of adolescent RBs in digital contexts, (2) to elucidate the mechanisms through which these behaviors emerge and impact well-being, and (3) to propose targeted, evidence-based interventions at the levels of practice, policy, and platform design. In doing so, this Perspective aims to bridge the gap between empirical research and actionable strategies, fostering safer and more equitable digital environments for adolescents worldwide. The following sections will first synthesize empirical findings, then discuss explanatory mechanisms, and finally explore integrated strategies for prevention, regulation, and design innovation. To situate the analysis, it is crucial to move beyond aggregate “screen time” metrics toward exposure- and experience-based measures that capture what adolescents actually do online (content type, interaction patterns, valence, and time-of-day), a direction strongly advocated by recent umbrella reviews and policy advisories [21,22,23]. This shift aligns with differential-susceptibility perspectives suggesting that individual traits and contexts moderate media effects, producing person-specific outcomes even under similar exposure conditions [5]. It also recognizes that “small average effects” at population level can translate into substantial public-health burden when exposures are ubiquitous and chronic, particularly for subgroups with compounding vulnerabilities [7,24].

This Perspective employs a structured narrative synthesis, drawing from peer-reviewed empirical and theoretical works published between 2020 and early 2025. Sources were identified through Scopus and PubMed searches using terms related to ‘adolescent risk’, ‘social media’, and ‘metaverse’, supplemented by manual screening of reference lists. Only studies addressing behavioral, cognitive, or affective mechanisms were included, ensuring conceptual coherence rather than exhaustiveness.

2. Synthesis of Evidence

Recent literature offers a detailed and often sobering account of the breadth of risky behaviors that adolescents engage in across social media and metaverse contexts, and the four articles provided by the author contribute important empirical depth to this picture.

2.1. Content Risks

Data from Piccoli et al. [25] highlight the interplay between peer group norms and cyberbullying, revealing that over 30% of surveyed adolescents reported engaging in online aggression in the past six months, with peer approval acting as a significant predictor. These findings align with broader global patterns showing gendered pathways: girls are more likely to encounter relational aggression in visual-heavy platforms like Instagram and TikTok, whereas boys more often face toxic peer interactions in competitive online gaming environments [26]. Other authors [27] also explored the link between body image concerns and exposure to idealized imagery on visual platforms, reporting that adolescents with high daily exposure had significantly higher odds of reporting dissatisfaction and disordered eating behaviors. These results complement recent research [28,29] while noting the moderating effect of active engagement with body-positive content, which shows potential for counteracting harmful norms [30]. Beyond these focal points, recent global analyses demonstrate that algorithmic recommender systems accelerate exposure to harmful content, such as self-harm communities and dangerous viral challenges, with minimal user engagement leading to rapid narrowing of suggested material toward extreme and risky themes [31,32].

2.2. Contact Risks

Patchin et al. [28] provide an in-depth analysis of grooming and sextortion cases, showing that more than 15% of adolescents surveyed had experienced unwanted sexual solicitations online, with financially motivated sextortion increasing sharply during the post-pandemic period. Europol points to the misuse of webcams, chat rooms in video games, social media, and encrypted messaging for grooming and sextortion, especially during the pandemic era; concurrently, NCMEC reports over 26,700 sextortion cases in 2024 (up from approximately 10,700 in 2022), with online enticement increasing by 192% [28,29]. In the realm of the metaverse, Di Pomponio and Cerniglia [9] found that VR harassment is perceived as more invasive and distressing than comparable experiences in 2D platforms, due to heightened presence and embodiment. Qualitative accounts from their participants describe persistent psychological aftereffects, which correlate with other studies emphasizing the emotional intensity of immersive interactions [33,34]. This adds a critical dimension to existing literature by illustrating how affordance-driven risks are amplified in embodied environments.

2.3. Conduct Risks

Problematic social media use, characterized by compulsive engagement and interference with daily functioning, is consistently associated with depression, anxiety, and sleep disturbances, with some longitudinal studies suggesting bidirectional relationships [35,36]. Similarly, early participation in seemingly benign “challenge” cultures can normalize boundary-pushing behaviors that offenders exploit for grooming and sextortion [28,29]. These conduct risks are further magnified in immersive contexts, where social validation cues and presence intensify affective arousal and risk-taking tendencies.

2.4. Economic and Structural Risks

Economic-behavioral risks such as loot box purchasing in games are consistently linked with problem gambling indicators and financial stress [37,38,39], raising urgent regulatory concerns. Finally, economic mechanics such as loot boxes embed variable rewards that intersect with adolescent reward sensitivity, with emerging evidence of migration from simulated to real-money gambling among a subset of youth [38,39,40]. These mechanisms reflect a convergence of economic design and developmental vulnerability, underscoring the need for regulatory attention to monetization systems targeting adolescents.

2.5. Interdependence of Risk Factors

Taken together, these findings—grounded in both the uploaded articles and complementary recent studies—underscore that while effect sizes are often small at the population level, the concentration of risk in vulnerable subgroups, combined with the compounding influence of platform design and socio-cultural context, warrants urgent attention from researchers, practitioners, and policymakers alike.

To deepen this synthesis, it is worth stressing the interdependence between content, contact, and conduct risks. Exposure to appearance-ideal content may co-occur with peer-sanctioned teasing, creating a feedback loop of social comparison and relational aggression [25,31]. In VR spaces, embodiment may heighten both the immediacy of harassment and the salience of pro-social support, implying that design and moderation choices can shift the balance between harm and help [9,12]. This interdependence extends to economic incentives as well, where monetized reward structures can reinforce both problematic engagement and susceptibility to social manipulation.

These interdependent risks collectively illustrate that adolescent vulnerability in digital environments cannot be attributed to isolated behaviors or single platforms. Rather, they emerge from the dynamic interplay between developmental processes, technological affordances, and social context. This complexity underscores the need for explanatory models capable of integrating neurocognitive, affective, and environmental mechanisms—a goal pursued in the following section.

3. Mechanisms and Explanatory Models

Understanding why adolescents are uniquely sensitive to online risks requires a multi-layered approach that integrates neurodevelopmental, technological, and socio-cultural perspectives. During adolescence, reward circuitry and socio-affective systems mature earlier than executive control regions, creating an imbalance between hyper-responsive socio-emotional systems—driven by limbic and striatal activity—and still-maturing prefrontal circuits responsible for self-regulation and executive control [1,3]. This developmental asynchrony heightens reward sensitivity and risk-taking tendencies, particularly in social contexts where peer evaluation is salient [4,5].

Digital environments amplify these vulnerabilities by providing constant social feedback, rapid reward cycles, and opportunities for peer comparison, intensifying the neurobiological drive toward engagement and experimentation. From an affordance-based perspective, platform features such as algorithmic personalization, infinite scroll, immersive embodiment, and ephemeral messaging lower the friction for risky actions and increase the salience of high-arousal content [13,30]. The metaverse adds unique affordances—spatialized presence, customizable avatars, and simulated physical proximity—which can heighten both positive social experiences and the perceived severity of negative encounters [9,12].

Socio-cultural moderators further shape how these mechanisms operate. Cultural norms regarding privacy, gender roles, and acceptable social conduct can either buffer or exacerbate risk. For example, collectivist orientations may intensify conformity to peer norms, magnifying the impact of online group dynamics on cyberbullying behavior [24]. Economic inequalities and access disparities also modulate exposure and vulnerability, with lower socio-economic status correlating with reduced digital literacy and fewer protective resources [7,10].

Crucially, these mechanisms are not additive but synergistic. A highly immersive VR environment emphasizing peer competition may simultaneously trigger neurobiological reward pathways, be reinforced by algorithmic surfacing of competitive content, and be intensified by cultural valorization of dominance behaviors. Recognizing these interdependencies is crucial for designing interventions that target the roots of risky behavior rather than its surface manifestations.

Moreover, person-specific susceptibility means that identical features can produce divergent outcomes—such as community belonging versus social comparison stress—depending on individual traits like baseline anxiety, prior victimization, or minority stress [22,41]. Future mechanism-focused studies should therefore employ high-resolution digital phenotyping, combining platform trace data and ecological momentary assessment, to map affective shifts to specific in-session events in immersive environments, where presence and embodiment dynamically modulate vulnerability [7,21].

4. Integrated Strategies for Prevention and Policy

Translating the synthesis of evidence and the explanatory models into actionable responses requires a coordinated, multi-level strategy. At the individual level, digital literacy and resilience programs tailored to adolescent developmental needs can mitigate vulnerability to risky behaviors. Evidence suggests that interventions focusing on critical consumption of online content, awareness of algorithmic influence, and peer-led education are particularly effective [42,43]. These should be integrated into formal curricula and supplemented by informal community-based initiatives. At the family level, parental mediation strategies—especially active mediation involving open discussions rather than restrictive monitoring—have been associated with lower incidence of cyberbullying perpetration and victimization [44]. For younger adolescents, co-use of immersive platforms under supervision may help model safe and respectful behavior while familiarizing caregivers with emerging technologies. In educational settings, schools should adopt whole-institution approaches that address online safety within broader well-being frameworks, combining policy enforcement, reporting mechanisms, and restorative justice practices [45]. Training educators to recognize signs of digital distress and to intervene appropriately in cases of online harassment or exposure to harmful content is essential. At the policy level, governments can strengthen regulations that mandate safety-by-design principles in platform development. This includes age-appropriate default settings, friction for high-risk features, transparent content moderation, and robust identity verification for high-contact environments [7,46]. International collaboration is critical for addressing cross-border risks such as grooming, sextortion, and illicit content distribution. From a design perspective, platforms and metaverse environments should implement proactive measures such as AI-driven detection of harmful behaviors, opt-in interaction models, personal safety boundaries in VR, and clearer consent protocols [9,30]. Incorporating adolescent user feedback into design processes through participatory co-creation can ensure that safety measures are both effective and acceptable to their intended audience. Finally, ongoing interdisciplinary research should inform all levels of intervention, ensuring adaptability as technologies and adolescent online cultures evolve. This includes longitudinal monitoring of risk trends, evaluation of intervention efficacy, and exploration of novel prevention avenues, such as integrating mental health support directly into platform interfaces [41]. The integration of these strategies across sectors offers the best prospect for reducing harm while preserving the developmental and social benefits of adolescent engagement in digital spaces. To strengthen implementation, we also recommend that platforms publish age-specific risk assessments and mitigation impact audits, analogous to safety cases in other regulated industries, and provide researcher access to privacy-preserving APIs for independent evaluation [7,47]. For monetization-related risks, probability disclosures for loot boxes, spending dashboards for guardians, and default spending caps can reduce harm without eliminating access altogether [40,47]. In immersive environments, default-on personal boundaries, proximity throttling, and one-click local muting/teleportation should be complemented by swift, transparent moderation responses and in-world evidence capture to support reporting and redress [12,30].

5. Conclusions and Future Directions

In conclusion, mitigating adolescent online risk requires moving beyond siloed approaches to adopt a unified strategy that embeds developmental science directly into platform design, governance, and education. Adolescent risky behaviors online arise from intertwined neurodevelopmental, socio-technical, and cultural factors, disproportionately harming vulnerable subgroups. Integrated, cross-sector strategies must anticipate evolving risks in immersive and AI-driven environments, engaging adolescents in design and governance. Research priorities include longitudinal, cross-cultural studies; real-time monitoring; and transparent data sharing. Interdisciplinary collaboration will be critical. Balancing connectivity benefits with harm reduction requires sustained, evidence-based, participatory approaches. Eight priorities for the next three years are: (1) preregistered audits of recommender pathways; (2) multi-arm trials manipulating affordances; (3) longitudinal gambling migration cohorts; (4) validated presence metrics and ethical VR experiments; (5) integrated sextortion surveillance with rapid response; (6) cross-cultural studies on narrative diffusion; (7) just-in-time adaptive safety interventions; and (8) standardized child impact assessments for high-risk features with public reporting [7,28,39,40].