The United Nations Children's Fund has sounded an alarm over the rapid pace at which young people are integrating artificial intelligence into their daily lives, far outpacing adult adoption rates. Speaking ahead of the inaugural Global Dialogue on AI Governance, UNICEF released findings from ten countries demonstrating that children are not only accessing AI systems at unprecedented speeds but are doing so with minimal protective guardrails in place. This technological shift represents a watershed moment in childhood development, one that carries profound implications for how societies must reshape their regulatory frameworks to protect vulnerable populations.
Data compiled by UNICEF indicates that at least 20 million children across the surveyed nations have already engaged with AI technologies. What distinguishes this adoption pattern from previous technological waves is both its velocity and the nature of the applications. Among these young users, approximately two million—representing roughly one in every ten child users—report actively turning to AI systems for personal counsel on matters causing them anxiety or distress. This represents a fundamental shift in how children seek guidance, moving away from traditional sources such as parents, educators, or mental health professionals toward algorithmic systems with unknown decision-making processes.
The educational dimension of AI adoption among children appears equally significant. An estimated 13 million young people utilise artificial intelligence to supplement their formal learning and complete homework assignments. This integration into academic workflows suggests that AI literacy is becoming embedded in childhood education whether or not schools have explicitly chosen to incorporate it. For many students, particularly in regions with limited access to tutoring or educational resources, AI systems may represent an equalising force. However, this benefit must be weighed against concerns about the quality of information provided and the potential for algorithmic bias to reinforce existing educational inequities.
UNICEF's analysis reveals a troubling asymmetry in the relationship between children and AI systems. Young people are extensively exposed to these technologies—experiencing firsthand how they function, what commercial interests drive them, and how personal data flows through their architecture—yet lack meaningful agency in avoiding them or challenging their operation. This exposure-versus-control imbalance becomes particularly acute when considering that children cannot easily opt out of systems embedded in schools, social media platforms, or entertainment applications that dominate their digital environment. The agency emphasises that children disproportionately experience the consequences of weak governance frameworks, yet will carry the long-term implications throughout their lives.
Concerns about harmful applications of AI among young users are widespread. According to the survey data, approximately one-third of children in the studied countries expressed worry about AI being weaponised for fraudulent purposes or to manufacture false information. These concerns reflect legitimate risks in the contemporary information landscape, where AI-generated content can seamlessly replicate authentic materials. An even more disturbing concern emerged regarding synthetic media manipulation: one-quarter of surveyed children reported anxiety about having their photographs or videos altered into sexually explicit deepfakes without consent. This particular fear speaks to the vulnerability of childhood innocence in an age where manipulative synthetic media can be generated with minimal technical expertise.
The absence of safety considerations in how many AI systems reach children represents a systemic failure of multiple stakeholders. UNICEF contends that too many platforms and applications deploy AI functionality without implementing robust safeguards, suggesting that protection has become an afterthought rather than a foundational design principle. This negligence stems partly from the speed of technological development outpacing regulatory capacity, but also reflects insufficient prioritisation of child welfare among technology companies and platform operators. The cultural assumption that innovation must proceed at maximum velocity, with safety addressed retrospectively if at all, proves particularly dangerous when the users are developmentally vulnerable populations.
Addressing this crisis demands comprehensive action across multiple institutional levels. UNICEF calls upon governments to embed child rights protections within international AI governance frameworks—a significant undertaking given that AI governance itself remains nascent and contested across jurisdictions. Financial investment in rigorous research examining how AI systems specifically harm children must increase substantially. Simultaneously, legal frameworks criminalising AI-enabled sexual exploitation require strengthening and harmonisation across borders, particularly given the transnational nature of digital harm. Private technology companies must commit to transparent AI design processes that prioritise safety, incorporating child protection experts in development stages rather than as compliance afterthoughts.
Building AI literacy among children themselves constitutes another essential component of protection strategies. Young people who understand both the capabilities and limitations of AI systems, as well as the business models driving their deployment, become more resilient against manipulation and more capable of recognising when their data is being exploited. This literacy must extend beyond technical understanding to encompass ethical reasoning about appropriate and inappropriate applications of AI. Yet building such literacy faces considerable challenges in regions where digital access remains limited or where educational systems lack capacity to incorporate emerging technological concepts.
The digital divide takes on new urgency in this AI context. Children in economically disadvantaged regions or countries with limited internet infrastructure may be excluded from AI-driven educational and economic opportunities, while simultaneously being exposed to AI-enabled fraud and exploitation. Ensuring equitable access to beneficial AI applications while protecting against harms requires targeted investment in digital infrastructure and education in underserved communities. For Southeast Asian nations, many of which are experiencing rapid digital development with varying regulatory capacities, this challenge looms particularly large.
UNICEF emphasises that decisions made about artificial intelligence governance today will reverberate across children's lives for decades, shaping their safety, privacy, wellbeing, and access to opportunities in a fundamentally AI-integrated world. This framing positions the current moment as genuinely consequential—not merely incremental regulatory adjustment but rather foundational choice-making about what kind of digital futures children will inherit. The stakes extend beyond individual protection to encompass broader questions about whether AI development will amplify or reduce existing inequalities, whether childhood will retain spaces free from algorithmic surveillance and manipulation, and whether young people will have meaningful voice in shaping technologies that profoundly affect them.
