Children are growing up with AI: Global AI Impact Summit can shape how

Artificial intelligence is reshaping childhood whether we intend it to or not. It’s helping millions of children learn better, access healthcare faster and navigate a changing world with greater opportunity. It could also deepen inequality, exclusion and risk at an unprecedented scale. The difference lies in the choices we make now about how AI is governed.

This week, as world leaders gather in Delhi for the Global AI Impact Summit, they have a critical opportunity to get this right. UNICEF is calling for an explicit recognition of children and young people in the Summit’s Leader’s Declaration as rights-holders, users and contributors to the AI ecosystem. This recognition would signal that AI governance is finally keeping pace with reality: Children are already at the centre of the AI revolution. 

AI is shaping childhood, but policy is lagging behind

Hundreds of millions of children are growing up in a world where AI shapes their education, entertainment, and opportunities. Yet governance frameworks have largely overlooked their specific needs and vulnerabilities. That must change.

We believe deeply in AI's potential to transform outcomes for children. It could personalise learning for students with disabilities, bring diagnostic tools to remote clinics, protect vulnerable communities from climate disasters, and deliver mental health support where it is needed most. But we must be honest about what we do not yet fully understand. The long-term impacts of sustained AI use on children's cognitive development, critical thinking, and mental health remain uncertain. Online risks including AI-generated child abuse material are evolving faster than our ability to address them.

AI set to redefine skills and youth employability

These uncertainties demand urgency, not caution. AI will radically reshape labour markets and force education systems worldwide to rethink what skills matter for youth employability. It may redefine education itself. We cannot be passive observers. We must govern this transformation actively and build the evidence base rigorously.

Child-centred AI means examining how systems affect children before deployment — their inclusion, access to services, data protection, and long-term development. It means designing for real-world conditions: offline access, shared devices, low-resource settings, social and language diversity. It means transparency, so children and caregivers understand when AI is being used and what recourse exists when harm occurs. And it means involving children meaningfully in design and governance.

Exclusion often comes from narrow design choices, a reality well understood by I-STEM's founder, Kartik, a visually impaired entrepreneur.

As a visually impaired entrepreneur, I've learned how exclusion can occur from narrow design choices. With support from the UNICEF Venture Fund, I-STEM built an AI accessibility platform that works across low-bandwidth contexts and multiple languages, enabling blind, low-vision, and dyslexic learners to access information others take for granted. We are piloting this with governments and partners in countries where UNICEF works. Designing with children strengthens innovation.

UNICEF partners govt, tech firms on child-centric AI policy

UNICEF is partnering with governments and technology companies to shape policies, test solutions, create child rights-based guidance for businesses, build evidence on what works and what risks emerge across education, health, child protection, and climate action. We are the partner of choice to help navigate the potential and risks of AI for this and coming generations. 

The Delhi Summit must commit to child-centred, age-appropriate, conducive and safe AI systems, underpinned by transparency, data protection, responsible business conduct, human oversight, and mechanisms for redress and remedy. Leaders must recognise that around one in three Internet users globally is under 18, yet two-thirds of school-aged children lack reliable home internet access. AI systems that ignore these realities risk hardwiring exclusion at an unprecedented scale.

Making this happen requires partnership, governments, industry, and civil society working not only for children, but with them. Decisions made in Delhi will echo across generations. We must act wisely. 

(Davin is Global Director, UNICEF, Office of Innovation; Sawhney is Founder, I-STEM. Views are personal.)