The digital divide between childhood and adulthood has taken on a new dimension as young people worldwide embrace artificial intelligence at a pace that far outstrips adult adoption. Speaking ahead of the first Global Dialogue on AI Governance, the United Nations Children's Fund revealed that children are incorporating AI into their daily lives at more than three times the rate of their parents and older peers, based on evidence gathered across 10 countries. This rapid uptake reflects how seamlessly AI has become woven into platforms and services that children naturally gravitate towards, from social media to educational apps, often without the critical awareness of the technologies shaping their online experiences.
The scale of this generational shift is staggering. UNICEF estimates that at least 20 million children globally have already engaged with AI systems in some form. What is particularly striking is the manner in which young people are turning to these tools for guidance on personal matters: more than two million children, representing roughly one in ten, have sought advice from AI on issues causing them worry or distress. This trend underscores both the appeal of AI as an accessible, non-judgmental information source and the corresponding gap in human support structures that might otherwise serve this function. For a generation growing up in an increasingly atomised digital landscape, AI has become an unexpected confidant.
Educational use represents another significant dimension of children's AI engagement. An estimated 13 million children worldwide are deploying AI to assist with their learning and homework completion. While this application might appear benign at first glance, it raises complex questions about academic integrity, critical thinking development, and the extent to which students are outsourcing cognitive tasks rather than developing their own problem-solving capacities. Schools and education systems across Southeast Asia and globally have struggled to establish coherent policies governing these tools, leaving educators and parents uncertain about appropriate boundaries.
The governance vacuum surrounding AI's reach into childhood represents the core concern articulated by UNICEF. Children exist in a paradoxical position: they are simultaneously the most exposed demographic to AI systems, directly experiencing their outputs and implications daily, yet they possess the least power to influence their design, understand the commercial incentives driving them, or opt out of their use. This power imbalance is particularly acute given that children have limited awareness of how their personal data is harvested, processed, and weaponised within AI training systems and business models. The agency emphasises that this generational exposure carries long-term consequences that children themselves did not choose and cannot easily escape.
Concerns about malicious applications of AI among young users are widespread and substantive. When surveyed across the ten countries studied, approximately one-third of child respondents expressed anxiety about AI being weaponised to perpetrate scams, manipulate peers, or manufacture and distribute misinformation. A quarter of children reported specific fear regarding non-consensual creation of deepfake imagery, particularly sexually explicit deepfakes. These concerns are not hypothetical: cases of AI-generated intimate imagery targeting young people have already surfaced globally, representing a serious extension of existing online harassment and exploitation patterns. For Malaysian and Southeast Asian families, these risks are compounded by the region's rapid digital adoption and, in some cases, limited regulatory oversight of online harms.
The underlying infrastructure problem persists despite growing awareness: too many AI systems continue to reach children with minimal built-in protections. UNICEF describes safety as appearing to be an afterthought rather than a foundational design principle. This is reflective of a broader industry dynamic where profit incentives and rapid deployment often supersede careful consideration of downstream harms, particularly those affecting vulnerable populations. The proliferation of AI applications in consumer technology means that children encounter these systems through mainstream platforms and services rather than specialist applications, making collective solutions difficult to implement.
UNICEF's prescribed response demands action across multiple institutional levels. Governments are called upon to embed child rights protections into international AI governance frameworks currently being negotiated and established. This includes legislative action specifically targeting AI-enabled sexual exploitation and abuse, as well as transparency requirements forcing companies to disclose how their systems function and how they utilise user data. The private sector bears responsibility for designing systems with safety and child welfare integrated from inception, not bolted on subsequently. Investment in rigorous research examining AI's specific risks and impacts on children remains inadequate, constraining evidence-based policymaking.
Digital literacy programmes represent another pillar of UNICEF's recommendations. As children become more sophisticated AI users, they require commensurate education in critical evaluation of AI outputs, understanding algorithmic bias, recognising manipulation, and safeguarding their personal information online. Southeast Asian countries with young, rapidly digitalising populations face particular urgency in implementing such programmes at scale. Simultaneously, the agency stresses the imperative to close the global digital divide: ensuring that AI literacy and safe access are not privileges of the wealthy, but available to marginalised communities where exposure to risks may be highest and protective resources lowest.
The timing of UNICEF's warning corresponds with a critical juncture in global AI governance. Regulatory frameworks are being constructed now that will determine the parameters within which these technologies develop and proliferate. The choices embedded in legislation, international standards, and industry practices established in the coming months will reverberate through childhood for decades. For Malaysia and the broader Southeast Asian region, this represents an opportunity to learn from experiences in other jurisdictions and to advocate proactively for child-centric governance rather than accepting frameworks designed elsewhere without sufficient consideration of local contexts, vulnerabilities, and values. The window for shaping AI's trajectory remains open, but it is closing with each new system deployment and each child user added to the ranks of early adopters.
