The United Nations Children's Fund has sounded a stark alarm about the accelerating adoption of artificial intelligence among young people, revealing that children worldwide are taking to the technology at rates more than three times faster than their adult counterparts. Drawing on fresh data from a decade of countries, UNICEF presented its findings at the inaugural Global Dialogue on AI Governance in Hamilton, Canada, underscoring the urgency of protecting vulnerable populations in an increasingly AI-driven world.
The scale of child engagement with these technologies is already substantial. UNICEF estimates that at least 20 million children across the surveyed nations have integrated AI into their daily lives, representing a significant and rapidly expanding cohort. Among this group, more than two million children—roughly one in every ten young AI users—have confided that they actively seek guidance from AI systems on matters causing them anxiety or distress. This reliance on algorithmic counsel for personal and emotional support raises fundamental questions about the nature of childhood development in the digital age and the adequacy of safeguards protecting minors.
The educational dimension of AI adoption proves equally pronounced. Approximately 13 million children report leveraging artificial intelligence to bolster their academic performance, whether through homework assistance, subject tutoring, or learning support. While such applications may promise educational benefits, the widespread integration of AI into schooling introduces variables that education systems and parents have had limited time to evaluate comprehensively. The interplay between AI-assisted learning and traditional pedagogical methods remains poorly understood, particularly regarding long-term cognitive and social development outcomes.
UNICEF's analysis reveals a troubling asymmetry between children's exposure to AI systems and their capacity to navigate or resist them meaningfully. Young people encounter these technologies with minimal awareness of their underlying design philosophies, the commercial incentives driving them, or how their personal data feeds into broader business ecosystems. Children possess neither the technical literacy nor the institutional power to challenge problematic AI deployments or opt out when systems prove harmful. This imbalance means that regulatory failures and governance gaps disproportionately impact those least equipped to protect themselves.
The security and reputational threats posed by AI technology weigh heavily on young users' minds. One-third of children surveyed across the ten nations expressed serious concerns about artificial intelligence being weaponised for deception—specifically its use in perpetrating scams or manufacturing false information designed to mislead. These anxieties are neither unfounded nor merely theoretical; documented cases of AI-enabled fraud and disinformation campaigns targeting young audiences have already emerged globally. The ability of AI to generate convincing false content at scale introduces a novel challenge to information literacy and trust in digital communications.
The threat of deepfake technology looms even larger in young people's consciousness. A quarter of children questioned reported anxiety about their images or videos being manipulated without consent to create sexually explicit synthetic content. This category of risk carries particular weight given the vulnerability of minors to sexual exploitation and the permanence of digital content. The psychological toll of knowing one's likeness could be weaponised in this manner extends beyond immediate victims to broader impacts on digital safety perceptions and online behaviour among youth populations.
UNICEF characterised the current landscape as one where technology systems reach children with inadequate protective measures in place. The characterisation of safety as an afterthought rather than a foundational design principle captures a systemic failure across technology industries and governance frameworks. Rather than embedding protections and ethical considerations from inception, many AI systems have been deployed with children as users or beneficiaries only after development cycles were substantially complete. This reactive rather than proactive approach to child safety amplifies risks and necessitates costly remediation efforts after harm has occurred.
For Southeast Asian nations including Malaysia, UNICEF's findings carry particular relevance given the region's rapid digital adoption and the relative nascence of comprehensive AI governance structures. Malaysia's young, tech-savvy population means local children likely mirror or exceed global adoption rates for AI technologies. The absence of robust domestic regulatory frameworks specifically addressing AI's impact on minors leaves Malaysian children potentially vulnerable to the same risks identified in UNICEF's international research, from data exploitation to deepfake abuse.
UNICEF has outlined a multi-stakeholder agenda for addressing these challenges. The organisation called upon governments, commercial entities, and civil society organisations to embed children's rights fundamentally into AI governance architecture. This encompasses substantial investment in research examining AI's specific risks to young users, strengthening legal protections against AI-facilitated sexual exploitation and abuse, mandating transparency and safety throughout the AI design process, expanding digital literacy education to help children navigate these tools critically, and bridging access divides to ensure marginalised children aren't left further behind.
The temporal dimension of AI governance decisions carries weight that extends far into the future. Current choices about how artificial intelligence develops, deploys to children, and remains regulated will reverberate through multiple generations. Young people encountering AI today will live longest with whatever safety gaps, privacy erosions, or developmental harms result from today's governance failures or successes. UNICEF's warning frames this not as a distant concern but as an immediate imperative demanding decisive action now to prevent entrenchment of harmful patterns.
Malaysia and other regional governments face mounting pressure to move beyond rhetorical commitments to child protection and establish concrete regulatory mechanisms governing AI's interaction with minors. This might encompass age-verification systems, parental consent requirements for data collection, mandatory safety audits before deployment to young users, and dedicated enforcement mechanisms with teeth. Without such frameworks, the global trajectory toward unguarded AI adoption among children will likely continue, amplifying risks that UNICEF has only begun to document comprehensively.
