The rapid integration of artificial intelligence into children's daily lives represents one of the most pressing digital governance challenges of our time. According to new research presented by the United Nations Children's Fund at the first Global Dialogue on AI Governance, young people are embracing AI technologies at a pace that dramatically outstrips adult adoption rates—more than three times faster based on data collected from 10 nations worldwide. This acceleration reveals a generation growing up with algorithmic systems as an assumed feature of their environment, yet largely without meaningful protections or understanding of how these tools operate.
UNICEF's findings paint a stark picture of the scale involved. The organisation estimates that at least 20 million children globally have already engaged with AI systems, a figure that continues to climb rapidly as these technologies become more accessible and integrated into educational platforms, social media, and consumer applications. Among this vast cohort, two million children—representing roughly one in every ten young users—have disclosed that they actively consult AI systems when facing problems or anxieties. This reliance on artificial intelligence as a source of counsel on matters of genuine personal concern underscores how seamlessly the technology has embedded itself in childhood experiences across diverse regions and economic contexts.
The educational dimension of AI adoption among children merits particular attention. An estimated 13 million young people now leverage artificial intelligence to facilitate their learning and complete homework assignments. While this application suggests potential academic benefits—from personalised tutoring to research assistance—it also creates dependencies that educators and parents may not fully comprehend. The proliferation of AI-powered learning tools without corresponding teacher training or curricular frameworks to address these technologies means children are navigating increasingly sophisticated educational systems without adequate guidance on how to evaluate information quality, recognise bias, or develop critical thinking skills.
What distinguishes UNICEF's alarm from routine concerns about technology adoption is the fundamental power imbalance it identifies. Children encounter AI systems with minimal agency to refuse, understand, or contest them. These young users are simultaneously subjects of data collection, targets of algorithmic manipulation, and consumers of services designed around business models that often prioritise engagement over welfare. The asymmetry extends deeper: children lack the legal standing to challenge how their personal information is harvested, processed, and monetised. They cannot negotiate terms of service or opt out of data practices that shape their digital experiences. Meanwhile, the adults designing and deploying these systems operate within governance frameworks that remain woefully inadequate across most jurisdictions.
The specific risks flagged by UNICEF paint a sobering picture of vulnerability. Approximately one-third of surveyed children across the 10 countries expressed concern about AI being weaponised for fraud, manipulation, and the dissemination of false information. A quarter reported anxiety about synthetic sexual imagery—deepfakes—being created using their own photographs or videos without consent. These are not hypothetical dangers. The technology enabling such harms already exists and circulates within platforms children access regularly. The gap between technological capability and protective regulation has created a climate where young people face sophisticated threats they are developmentally unprepared to navigate.
The governance failure extends beyond individual harm to encompass systemic issues. UNICEF emphasises that numerous AI systems reach children with minimal safety mechanisms in place, suggesting that child protection has functioned as an afterthought in the design and deployment cycle rather than a foundational principle. This reflects broader patterns in the global AI industry, where speed to market and profit maximisation have consistently outpaced ethical review. Unlike traditional consumer products that undergo safety testing before reaching young users, many AI applications deploy at scale and iterate based on user feedback—with children serving as uncompensated test subjects exposed to unknown risks.
For Southeast Asian nations including Malaysia, these international trends carry particular significance. The region has witnessed explosive growth in smartphone penetration and internet connectivity, with children representing a substantial portion of new digital users. Malaysian families, like their counterparts across Asia, increasingly rely on AI-powered educational platforms and online services. Yet Malaysia's regulatory framework for AI remains nascent, with no comprehensive legislative approach specifically protecting children from algorithmic harms. The country risks importing the governance gaps UNICEF identifies while lacking the domestic capacity to address them through local solutions attuned to Malaysian contexts and values.
UNICEF's prescription for change encompasses multiple levels of intervention and responsibility. Governments must embed child rights protections directly into emerging AI governance structures rather than treating them as secondary considerations. This requires substantial investment in research specifically examining how AI systems affect children's development, safety, and opportunities—a knowledge gap that currently hampers evidence-based policymaking. Legislative frameworks must evolve to criminalise and effectively prosecute AI-enabled sexual exploitation and abuse, closing loopholes that currently allow perpetrators to operate with impunity. Private sector actors designing and deploying AI systems must commit to transparent development processes and genuine safety-first design principles rather than cosmetic compliance.
Addressing the digital divide represents another crucial dimension. As AI systems increasingly mediate access to education, information, and economic opportunity, children without reliable internet access or devices capable of running these applications face compounding disadvantage. The promise of AI to democratise knowledge rings hollow when access remains stratified by geography and economic status. Building genuine AI literacy among young people—understanding not just how to use these tools but how they function, what data they collect, and what biases they may encode—requires educational investment that most nations have scarcely begun.
The moment UNICEF describes as decisive captures a genuine inflection point. Decisions made now about how AI governance frameworks emerge will shape childhood experiences for generations. If current trajectories continue unchecked, an entire cohort of young people will grow to adulthood having normalised surveillance, accepted algorithmic decision-making, and internalised the notion that their data and attention represent commodities to be harvested. Conversely, if governments and industry respond with genuine commitment to protection and transparency, childhood in the AI era could be fundamentally different. For Malaysia and other developing nations in the region, this represents an opportunity to learn from early mistakes in other jurisdictions and build AI ecosystems that centre child welfare from inception rather than attempting remedial fixes after harms have proliferated.
