The United Nations Children's Fund has sounded a stark warning about the accelerating pace at which young people are embracing artificial intelligence technologies, outpacing adult adoption by more than threefold according to fresh research spanning 10 countries. Speaking ahead of the inaugural Global Dialogue on AI Governance, UNICEF emphasised that while AI systems are becoming an inescapable part of childhood experiences worldwide, the consequences—both positive and concerning—remain largely uncontrolled and inadequately understood.
The scale of children's engagement with AI is substantial and growing rapidly. UNICEF's data indicates that at least 20 million children globally have already used artificial intelligence systems in some form, demonstrating how thoroughly these technologies have penetrated young people's daily lives across education, entertainment, and social interaction. More troublingly, over two million children—representing approximately one in every 10—have acknowledged turning to AI systems for guidance on matters that cause them anxiety or distress, effectively treating algorithms as confidants for their personal concerns.
Educational applications represent another major domain where children are integrating AI into their routines. The agency estimates that roughly 13 million children rely on AI tools to supplement their learning processes and complete homework assignments. This trend reflects a broader shift in how knowledge acquisition is occurring, with algorithms increasingly mediating the learning experience. While such applications can enhance educational outcomes, they simultaneously create dependencies and raise questions about critical thinking development and the quality of information these systems provide to developing minds.
What distinguishes children's relationship with AI from adults' is the profound asymmetry of power and agency. Young users encounter these systems with minimal understanding of their underlying mechanisms, the commercial interests driving their design, or how their personal data is harvested and monetised. Children possess far fewer tools to recognise manipulation, avoid problematic content, or challenge algorithmic decisions that affect them. This vulnerability becomes particularly acute given that the long-term consequences of early AI exposure will unfold across their entire lifespans, yet they have negligible say in governance decisions shaping these technologies.
The risks children face extend across multiple dimensions. One-third of the 10,000-plus children surveyed reported anxiety about AI being weaponised to perpetrate fraud, deceive vulnerable individuals, or facilitate the rapid spread of false information. In an era of sophistication deepfake technology, a quarter of respondents expressed specific fears about having their images or videos manipulated into sexually explicit content without consent. These are not hypothetical dangers but emerging realities already documented in schools and communities across multiple continents, including Southeast Asia where such incidents have surfaced in recent years.
The current regulatory landscape offers inadequate protection. UNICEF characterises the situation as one where numerous AI systems are reaching children entirely unguarded, with child safety apparently relegated to an afterthought in design and deployment decisions. This governance vacuum means children are effectively serving as test subjects for technologies that remain poorly understood even by their creators, with vulnerable populations bearing disproportionate risks while corporations and governments benefit from the data harvest and behavioural insights they generate.
For Malaysian and broader Southeast Asian contexts, these findings carry particular urgency. The region's young population is digitally native and increasingly integrated into global AI ecosystems, yet most countries lack comprehensive legislative frameworks addressing AI-specific harms to children. The rapid proliferation of AI chatbots, content recommendation systems, and educational platforms has outpaced policy development, leaving young users navigating these environments with minimal institutional protection. Local content moderation remains inadequate, and the digital divide means rural and lower-income children face both greater vulnerability and fewer quality alternatives.
UNICEF's recommendations call for fundamental reorientation of how societies approach AI governance. Governments must embed child rights protections at the core of AI policy development rather than treating them as supplementary considerations. Investment in research specifically examining AI's impacts on children's psychological development, privacy, and safety remains critically underfunded. Legal frameworks must be substantially strengthened, particularly regarding AI-enabled sexual exploitation, which currently operates in many jurisdictions with minimal enforcement capacity.
Transparency and design ethics must become central to AI system development. The agency advocates for mandatory safety assessments before child-accessible systems launch, clear disclosure of how algorithms personalise content, and genuine mechanisms for children and parents to understand and contest decisions affecting them. Critically, AI literacy programmes must equip young people with realistic understanding of these technologies—neither dystopian fear nor uncritical acceptance—enabling them to navigate digital environments with greater autonomy.
Addressing the digital divide also emerges as essential. While affluent children in developed nations may access higher-quality, better-moderated AI applications and receive stronger digital education, children in less developed regions often encounter lower-quality, more exploitative systems. This technological inequality risks widening existing socioeconomic disparities, concentrating the harms of inadequately governed AI among the world's most vulnerable young populations.
The window for action, UNICEF emphasises, remains open but rapidly closing. Current decisions about AI governance will establish precedents and trajectories influencing children's safety, privacy, and opportunity for decades forward. Without immediate, comprehensive action involving governments, technology companies, civil society, and international bodies, the trajectory suggests children's exposure to AI risks will continue escalating while safeguards lag further behind. For Southeast Asian policymakers, the message is clear: intentional, child-centred AI governance is no longer optional but essential to protecting future generations.
