The United Nations Children's Fund has raised urgent alarms about the rapid proliferation of artificial intelligence among children, revealing that young people are embracing AI technologies at a pace dramatically outstripping adult adoption rates. Drawing on data from ten countries, UNICEF highlighted in a statement released ahead of the inaugural Global Dialogue on AI Governance that children are now front-and-centre in the digital revolution, yet remain dangerously vulnerable to its risks.
The scale of AI penetration among the young is staggering. UNICEF estimates that at least 20 million children globally have already engaged with AI systems of some form. More troubling still, the agency found that more than two million young users—roughly one in ten of those surveyed—are turning to AI platforms for guidance on personal matters that cause them worry or distress. This represents a fundamental shift in how children seek counsel and reassurance, with algorithmic systems increasingly replacing traditional sources of advice such as parents, teachers, or counsellors.
The educational dimension of AI adoption among youth is substantial. Approximately 13 million children across the surveyed nations reported deploying AI to enhance their academic performance and complete homework assignments. While this suggests potential benefits for learning, it also underscores a broader trend: AI is becoming deeply embedded in children's daily routines and developmental processes, often without adequate oversight or consideration of long-term implications.
What distinguishes UNICEF's concerns from typical technology anxiety is the systematic nature of the vulnerabilities. Children face exposure to AI systems whose design architectures, profit-driven business models, and data harvesting practices remain opaque to them. The organisation emphasises a critical asymmetry: young people experience the full force of AI technologies yet possess minimal agency to decline participation or push back against their integration into their lives. This power imbalance is particularly acute given that children lack the developmental maturity, legal standing, and technological literacy to advocate for their own interests.
The governance vacuum leaves children exposed to multiple categories of harm. A third of children in the ten-country sample expressed anxiety about AI being weaponised for fraud and deception, or deployed to manufacture and distribute false information. These concerns are not theoretical—they reflect real vulnerabilities in an ecosystem where AI-generated scams and misinformation campaigns increasingly target younger, less discerning audiences. The psychological and social consequences of exposure to such content during formative years remain poorly understood.
Sexual exploitation represents perhaps the most sinister threat. A quarter of surveyed children reported apprehension about the creation of non-consensual deepfake imagery depicting them in sexually explicit scenarios. This fear is grounded in documented cases of such abuse, where AI technology has enabled perpetrators to manufacture synthetic abuse material at scale and with minimal technical skill. For children facing such threats, the violation is profound, combining elements of image-based abuse with the permanence and spreadability of digital content.
UNICEF's diagnosis of the problem is unsparing: safety mechanisms are conspicuously absent from systems being deployed to children, with corporate priority apparently given to product rollout over protective guardrails. The agency's statement explicitly notes that safety considerations appear to be an afterthought in the development and deployment pipeline. This reflects broader industry dynamics where growth and market capture typically outweigh protective measures, particularly where vulnerable populations are concerned.
For Malaysian and Southeast Asian readers, the implications are locally acute. The region is experiencing rapid AI adoption driven by tech companies seeking to expand beyond saturated Western markets, yet regulatory frameworks remain underdeveloped. Child protection standards in AI governance are virtually non-existent across much of the region, and parental awareness of AI-specific harms remains limited. Local content moderation capacity is inadequate, meaning AI-generated scams and deepfakes circulating in regional languages often evade detection and removal mechanisms designed for English-language content.
UNICEF's prescriptive recommendations target multiple actors simultaneously. Governments are urged to prioritise research illuminating how AI systems specifically harm children, to criminalise AI-enabled sexual abuse more comprehensively, and to enforce transparent design standards that allow independent safety audits. The private sector is called upon to embed child protection from inception rather than retrofitting safety measures. Broader initiatives must focus on equipping young people with AI literacy skills that enable them to navigate these systems critically, while simultaneously addressing the digital divide that leaves disadvantaged children unable to access protective technologies or quality AI educational resources.
The timing of UNICEF's intervention reflects a narrowing window for preventative action. AI governance structures are currently being established globally, with many nations and international bodies crafting initial policy frameworks. The decisions made at this juncture regarding child protection, data rights, and AI accountability will reverberate for decades, shaping whether generations of young people grow up in an ecosystem with meaningful safeguards or one characterised by exploitation and risk. For Southeast Asia particularly, where regulatory institutions are still developing and tech company influence is substantial, the choices made now will determine whether regional youth face the same governance vacuum or whether child-centred AI principles take root from the outset.
