The United Nations Children's Fund has sounded an urgent alarm about the rapid integration of artificial intelligence into children's daily lives, revealing that young people are adopting these technologies at a pace substantially outstripping adult uptake. Speaking ahead of the inaugural Global Dialogue on AI Governance, UNICEF disclosed findings from research spanning 10 countries that paint a picture of a digital generation increasingly reliant on AI systems with minimal protective oversight. The scale of this shift is striking: at least 20 million children globally have already turned to artificial intelligence, a figure that underscores how quickly these tools have become embedded in youth culture and behaviour.
The scope of children's engagement with AI extends well beyond casual technological curiosity. More than two million youngsters surveyed—representing approximately one in ten—have begun consulting artificial intelligence when facing personal worries or uncertainties, effectively treating these systems as trusted advisors on matters that might previously have prompted conversations with parents, teachers, or peers. This fundamental change in how children seek guidance carries profound implications for child development, psychological wellbeing, and the quality of information available to vulnerable minds. The pattern becomes even more pronounced when examining educational usage: an estimated 13 million children across the surveyed nations have incorporated AI into their learning routines, relying on these systems to complete homework assignments and support their academic development. For Malaysian educators and policymakers watching these trends, the shift signals both unprecedented educational potential and serious questions about how local students are being prepared to navigate AI-mediated learning.
What distinguishes UNICEF's concerns from typical technology adoption concerns is the fundamental asymmetry of power and knowledge it identifies. Young users possess minimal understanding of how artificial intelligence systems operate, what commercial motivations drive their design, or how their personal data feeds into training algorithms and business models. Children encounter these systems with almost no ability to refuse them, question their premises, or protect themselves from their inherent risks. The agency argues persuasively that this vulnerability gap—where children are disproportionately exposed to AI's mechanisms yet possess disproportionately little power to challenge or circumvent them—represents a critical governance failure. In the Southeast Asian context, where rapid digital adoption often outpaces regulatory development, this observation carries particular weight.
The specific harms children face through inadequately governed AI systems paint a sobering picture. Approximately one-third of children in the 10 countries examined expressed anxiety about artificial intelligence being weaponised to perpetrate scams, manipulate peers, or manufacture false information. An equally troubling one-quarter reported fears centred on a particularly sinister risk: the creation of sexually explicit deepfakes using their own images or videos. These are not theoretical concerns; they represent real harms already documented in jurisdictions worldwide, from synthetic non-consensual intimate imagery to elaborate romance scams targeting teenagers. The psychological burden of knowing such technological violations remain possible creates a distinct form of digital vulnerability unique to this generation.
UNICEF's characterisation of the current regulatory landscape as permitting systems to reach children "with no guardrails" reflects a stark reality: the commercial incentives driving AI deployment have far outpaced the development of protective frameworks. Safety considerations appear to function as afterthoughts rather than fundamental design principles, and the burden of protection has fallen almost entirely on parents and institutions rather than being embedded into technological architecture itself. This approach proves manifestly inadequate for children who encounter AI systems in schools, through peer networks, via social media platforms, and through educational applications designed with minimal consideration of age-appropriate safeguarding. For Malaysian parents and administrators, this underscores how global technological governance failures directly impact local children's safety.
The organisation has articulated a comprehensive agenda for addressing these deficiencies, calling on multiple stakeholders—governments, private technology companies, and civil society partners—to fundamentally reorient how artificial intelligence is governed with children's interests as a central consideration. This agenda extends across multiple dimensions: substantial investment in research specifically examining how artificial intelligence threatens children; strengthened legislative frameworks targeting AI-facilitated sexual exploitation; design processes that prioritize safety alongside transparency; systematic education to build young people's capacity to understand and critically engage with artificial intelligence; and deliberate efforts to ensure digital access gaps do not compound other inequalities. Each component recognises that effective protection requires not isolated interventions but rather systemic change across multiple institutions and sectors.
For Malaysia and its Southeast Asian neighbours, the timing of this intervention proves particularly significant. Many nations in the region are only beginning to develop comprehensive digital governance frameworks, presenting an opportunity to embed child protection principles from the outset rather than retrofitting safeguards to existing systems. The potential to shape how AI governance develops in the regional context—before dominant practices become entrenched—could position Southeast Asian nations as leaders in child-centred digital policy. Conversely, proceeding without adequate safeguards risks replicating mistakes already evident in more developed technology markets, where protections remain fragmentary and reactive.
UNICEF frames the current moment as decisional in character: the choices made now about artificial intelligence development, deployment, and governance will fundamentally determine what opportunities and what harms children experience for decades to come. The trajectory established through present governance decisions will shape whether children's relationship with artificial intelligence becomes one characterised by agency and safety or by vulnerability and manipulation. This intergenerational dimension gives unusual weight to what might otherwise seem like technical policy questions. The organisation emphasises that delaying action effectively represents choosing to expose increasingly large cohorts of young people to inadequately managed risks during the most formative years of their development.
The data and advocacy UNICEF has presented create pressure on policymakers to move beyond incremental measures toward comprehensive reform. For Malaysian stakeholders in government, education, technology, and civil society, this represents both challenge and opportunity: the challenge of developing protective frameworks quickly enough to address harms already materialising, and the opportunity to shape how a transformative technology relates to the generation for whom it will form the background of ordinary life. Without coordinated action that embeds child rights as a central consideration in artificial intelligence governance, the organisation suggests, the default trajectory will continue privileging commercial innovation over vulnerability protection—a path that will reserve both AI's benefits and its harms differentially across childhood populations.
