The public and media discourse surrounding children, adolescents, and the internet is frequently caught in a loop of adult-centric alarmism, or what sociology defines as “moral panic” (Cohen, 1972; Finkelhor, 2011). Contemporary youth are often labeled pejoratively as the “Generación de Cristal” (Glass Generation), presenting them as an exceptionally fragile, hyper-sensitive, and easily manipulated cohort that lacks resilience when navigating the pressures of the modern digital landscape (Gentil González, 2026). This paternalistic perspective often frames the internet as a pure “risk amplifier” that inevitably leads to isolation, radicalization in the Manosphere, or exposure to complex political disinformation campaigns (Gentil González, 2026).

However, empirical research reveals a massive disconnect between adult cultural anxieties and the actual daily practices of youth. A recent Danish study by Adler Berg et al. (2026), a mixed-methods investigation of 5- to 12-year-olds published in New Media & Society, demonstrates that children do not actively seek out or consume digital spaces through an ideological lens. Instead, they experience online deception as an extension of their play culture (Adler Berg et al., 2026; Boyd, 2014). The vast majority of what children perceive as “misinformation” consists of funny, absurd videos (40%), such as a viral hoax claiming sea foam is rich in vitamins because it is “whale pee” (Adler Berg et al., 2026). When they do encounter real deception, it is highly pragmatic and ludic: fake tips, tricks, and “scams” (such as fake generator links promising “free Robux” in Roblox or skins in Fortnite) (Adler Berg et al., 2026). In fact, 24% of children aged 10–12 report encountering such gameplay deceptions (Adler Berg et al., 2026). For the youngest internet users, the digital world is a playground where lies and exaggerations are often shared simply “for fun,” leaving little negative emotional footprint (Adler Berg et al., 2026).

How do we bridge this gap? How do we safeguard children from structural online dangers without resorting to clumsy, paternalistic prohibitions that alienate the very people they are meant to protect?

The solution lies in replacing moralistic lectures with applied science—specifically, explainable artificial intelligence (XAI), privacy engineering, and robust cybersecurity education. This principle largely synthesizes the research focus and project work of the Cybersecurity and Privacy Protection Research Group (GiCP) at the Spanish National Research Council (CSIC - ITEFI).

The Algorithmic “Trojan Horse”: Exposing the Affinity Funnel

While children engage with the internet playfully, professional disinformation campaigns and malicious actors exploit this exact play culture to propagate harmful content (Adler Berg et al., 2026; Olmos Alcaraz, 2025). This is the “Trojan Horse” effect:

  1. The Hook: A child searching for gaming cheats or free virtual currency is lured out of regulated platforms and into closed, third-party messaging servers (like custom Discord channels) (Adler Berg et al., 2026).
  2. The Smuggling: Inside these private, peer-to-peer spaces, radicalized narratives, misogynistic tropes, or xenophobic hoaxes—such as the systematic dehumanization of unaccompanied migrant minors (under the acronym MENA)—are introduced under the guise of “edgy” irony and dark humor (Olmos Alcaraz, 2025).
  3. The Escalation: Once the child interacts with this “edgy” content, the platform’s recommendation algorithms register high engagement. The child is then funneled into an “affinity funnel,” where their feed is progressively microsegmented, bombarding them with targeted radical or pseudoscientific content (such as extreme muscle-building advice or diet scams linked to the Manosphere) (Gentil González, 2026).

To counter this invisible funnel, the R&D project XAI-Disinfodemics. The project developed a socio-technological methodology that analyzes multimodal disinformation (incorporating text, images, and stylistic patterns) to provide explainability (Explainable AI - XAI). Rather than simply labeling a post as “fake,” this methodology explains how and why a piece of media is manipulative. In a classroom setting, this provides teachers with a tool for relational scaffolding (inspired by Vygotskian child development theory) (Boyd, 2014), moving away from blind prohibition toward teaching students to deconstruct the technical mechanics of algorithmic manipulation.

SafeHorizon: Disrupting the Cybercrime Bazaar

The deception children face online is not merely cognitive; it is structurally tied to the global underground economy. When an adolescent downloads a fake Fortnite cheat or a cracked game and disables their antivirus to run it, they aren’t just trying to win—they are unknowingly executing infostealers and trojans designed to steal the family’s personal credentials (SafeHorizon Consortium, 2024).

The GiCP actively addresses this threat on an international scale through its involvement in the European project SafeHorizon (Horizon Europe, Grant Agreement no. 101168562, running from 2024 to 2027) (SafeHorizon Consortium, 2024). SafeHorizon focuses on disruptively detecting “Crime-as-a-Service” (CaaS) and Malware-as-a-Service (MaaS) operations. By analyzing Dark Web marketplaces, credential-trafficking networks (such as traffers selling stolen clouds), and spoofing attacks the project maps the actual infrastructure that powers digital deception.

Crucially, under CSIC’s ethical leadership within the consortium, SafeHorizon ensures that these investigative tools respect fundamental rights, data protection, and privacy laws (SafeHorizon Consortium, 2024). The GiCP’s technical developments—such as open-source DNS validation tools and group-signature cryptographic protocols—prove that digital defense requires hard engineering, not moral scolding. Educating youth about this “cybercrime bazaar” demystifies the “cool” factor of hacking, showing them that a “harmless cheat” is actually fuel for transnational criminal syndicates.

The Folly of Blanket Bans: Lessons from Australia’s SMMAR

Faced with these digital threats, governments worldwide often default to blanket prohibitions, such as banning social media for anyone under the age of 16 (Petersen et al., 2026; SMC Spain, 2026). However, scientific evaluations of these measures reveal that legal bans are highly ineffective and often counterproductive.

A landmark study published in JAMA Network Open (August 2026) assessed the impact of Australia’s Social Media Minimum Age Restriction (SMMAR) and teenagers’ responses to it (Petersen et al., 2026). Evaluated and disseminated by the Science Media Centre (SMC) Spain, the findings and expert commentaries highlight several critical flaws in the “blanket ban” approach (SMC Spain, 2026):

• Circumvention by Default: The study found that 75% of teenagers explicitly planned to continue using social media by finding technical ways to bypass the restrictions (Petersen et al., 2026; SMC Spain, 2026).

• Creating Less Safe Spaces: As Professor Tatiana Íñiguez Berrozpe (University of Zaragoza) warns, legally banning platforms does not stop youth from using them; it merely pushes them into unmonitored digital spaces with less parental oversight, while failing to protect vulnerable groups who rely on these networks for social connection, identity exploration, and support (SMC Spain, 2026).

• Legal Restrictions Do Not Solve Well-Being: Professor María del Mar Sánchez Vera (University of Murcia) points out that simply raising the access age does not address teenagers’ well-being, as social media use is driven by deep social needs for peer connection and entertainment rather than mental health pathology (SMC Spain, 2026).

• Weak Mental Health Correlations: Professor José César Perales (University of Granada) emphasizes that standard scientific literature consistently shows extremely weak or clinically insignificant correlations between social media screen time and mental health indicators like anxiety or depression (SMC Spain, 2026). These minor associations are heavily confounded by major socio-demographic factors—such as socioeconomic status, family dynamics, and physical support networks—which are often ignored in alarmist headlines (SMC Spain, 2026).

Privacy Engineering as the Real Firewall

If blanket bans fail, and if the “Trojan Horse” relies on microsegmenting children based on their psychological profiles, then data privacy is not a bureaucratic chore—it is the ultimate firewall for youth mental health and autonomy.

To “starve” recommendation algorithms of the behavioral data they use to target adolescents, a conceptual policy framework could propose the implementation of school-level privacy engineering metrics (designated by their Spanish acronyms):

• IREFE (Índice de Resiliencia del Entorno Familiar y Escolar / Family-School Digital Resiliency Index): A proposed metric tracking the speed at which Spain’s Canal Prioritario (AEPD) executes urgent injunctions to remove non-consensual humiliating content, minimizing the viral exposure of digital violence.

• IICE (Índice de Integridad de Canales Escolares / School Trust Channel Integrity Index): A proposed metric to evaluate the enforcement of Multi-Factor Authentication (MFA) and secure VPNs in teacher teleworking. This prevents attackers from hijacking trusted school communications to distribute highly credible hoaxes and phishing scams to families.

• ICMP (Índice de Control de Microsegmentación y Perfilado del Alumnado / Student Profile and Microsegmentation Control Index): A proposed auditing standard for school-issued devices (Chromebooks, tablets) to ensure they block third-party cookies and behavioral tracking by default, breaking the loop of targeted algorithmic radicalization.

Conclusion: From Technopanic to Dialogue

Protecting the next generation requires us to abandon “moral panics” and adult-centric paternalism (Cohen, 1972; Finkelhor, 2011). Instead of implementing ineffective bans that push children into the shadows (SMC Spain, 2026), we must establish an “epistemic horizontality”—a dialogic, intergenerational relationship based on mutual trust and shared digital literacy (Boyd, 2014).

By combining peer-led active citizenship (such as the Cibercorresponsales and its “La Pinza” campaign) with the rigorous technical standards of explainable AI and privacy engineering demonstrated by groups like the GiCP-CSIC, we can co-create a digital ecosystem where youth are neither fragile victims nor lawless actors, but secure, critical, and autonomous citizens.

References

  1. Adler Berg, F. S., Heiselberg, L., Lundtofte, T. E., & Frischlich, L. (2026). Growing up with misinformation: A mixed-methods study of the perceptions, experiences, and responses among the 5- to 12-year-olds. New Media & Society, 28(1), 1–25. https://doi.org/10.1177/14614448261435124
  2. XAI-Disinfodemics Consortium (2021-2024). XAI: Explainable AI for Disinformation and Conspiracy Detection During Infodemics. Agencia Estatal de Investigación (AEI). https://www.aei.gob.es/noticias/xai-disinfodemics-ia-analiza-textos-imagenes-emociones-frenar-desinformacion-polarizacion
  3. Boyd, D. (2014). It’s Complicated: The Social Lives of Networked Teens. Yale University Press.
  4. Cohen, S. (1972). Folk Devils and Moral Panics: The Creation of the Mods and Rockers. Routledge.
  5. Finkelhor, D. (2011). The Internet, Youth Safety, and the Problem of “Juvenoia”. Crimes Against Children Research Center, University of New Hampshire.
  6. Gentil González, C. (2026). Más allá del pánico moral: radiografía del descontento juvenil y el silencio de las aulas. Análisis Nuevo Diagnóstico. https://nuevodiagnostico.org/radiografia-descontento-juvenil
  7. Olmos Alcaraz, A. (2025). Disinformation, Xenophobia, and the Dehumanisation of Unaccompanied Migrant Minors (MENA) in Spain. Journal of Youth and Migration Studies, 14(2), 45–68.
  8. Petersen, M., et al. (2026). Evaluation of Adolescent Intention to Circumvent Minimum Age Restrictions on Social Media in Australia. JAMA Network Open, 9(8), e2624100. https://doi.org/10.1001/jamanetworkopen.2026.24100
  9. SafeHorizon Consortium. (2024). SafeHorizon: Innovations in Detecting and Disrupting Crime-as-a-Service Operations. Horizon Europe Grant Agreement No. 101168562. SafeHorizon
  10. Science Media Centre (SMC) Spain. (2026). Most young Australians intended to continue using social media, even if the ban came into force. Expert Reactions on Petersen et al. (2026) in JAMA Network Open. https://sciencemediacentre.es/en/most-young-australians-intended-continue-using-social-media-even-if-ban-came-force