Reading UNICEF’s Guidance on AI and Children as Public Infrastructure

Feb 6
3 min read

UNICEF Innocenti’s Guidance on AI and Children provides one of the most developed frameworks to date for understanding how artificial intelligence affects children’s rights, agency, and life chances.

The document treats children not as abstract future users of technology, but as people whose lives are already shaped by algorithmic systems. Education, welfare, health, protection, migration, and content moderation systems increasingly rely on AI. The guidance recognizes this reality and addresses AI as a structural force in children’s environments.

This framing matters. It places AI governance within the domain of rights, institutions, and long-term responsibility.

How the guidance structures responsibility

The document organizes AI governance around children’s rights as defined in the UN Convention on the Rights of the Child. It maps how AI systems intersect with rights to privacy, non-discrimination, participation, protection, and development.

Governance is treated as a continuous obligation across the AI lifecycle. Design choices, data practices, deployment contexts, and monitoring mechanisms are all addressed. The guidance assigns responsibility to states, companies, educators, and caregivers, recognizing that AI systems operate across multiple domains simultaneously.

This approach establishes AI governance as a shared societal responsibility rather than a technical specialization.

What the guidance makes clear

Several elements stand out.

First, children experience AI asymmetrically. They are more exposed to automated decision-making and less able to contest it. The guidance recognizes this imbalance and treats it as a governance problem rather than an unfortunate side effect.

Second, participation is treated as a right. The document emphasizes that children have a stake in systems that shape their lives and that their perspectives must inform design and policy. This frames participation as part of governance, not outreach.

Third, long-term impact is central. The guidance focuses on developmental consequences over time, including how early exposure to algorithmic systems shapes autonomy, opportunity, and social inclusion.

Together, these elements position AI as part of children’s lived environment and therefore subject to public accountability.

What this means for public institutions

For public institutions, this guidance establishes clear expectations.

AI systems that affect children require heightened governance. Risk assessment, transparency, and accountability must reflect children’s limited ability to exit, contest, or negotiate system use.

Institutions responsible for education, welfare, health, and protection carry direct responsibility for how AI systems shape children’s outcomes. Governance cannot be delegated entirely to vendors or framed solely as compliance.

The UNICEF guidance provides a rights-based foundation for institutional responsibility in these domains.

How federated governance extends this work

As AI systems affecting children increasingly operate across platforms, jurisdictions, and sectors, governance also needs to operate across institutional boundaries.

Federated governance provides that structure.

It enables shared oversight of systems used by schools, social services, platforms, and public agencies. It distributes authority so that no single actor determines how children’s data, profiles, or opportunities are shaped.

Federated governance supports children’s rights through concrete mechanisms:

Exit, so institutions can move away from systems that undermine rights without losing capacity
Voice, so children, caregivers, and advocates have standing in governance processes
Transparency, so system behavior and changes are visible and contestable

These mechanisms translate rights into enforceable structure.

Why this connection matters

The UNICEF guidance establishes a strong normative and institutional foundation. It defines what responsible governance requires when children are affected by AI systems.

Federated governance provides the infrastructure layer that allows those responsibilities to persist across systems, providers, and political change.

Together, they form a model of AI governance that treats children’s environments as public infrastructure and children themselves as rights-bearing participants in its design and oversight.

The UNICEF Innocenti guidance contributes a critical piece of this architecture. It clarifies what is at stake. It names who carries responsibility. It frames AI governance as a long-term obligation to those with the least power and the most to lose.

Why Mission AI is sharing this

Mission AI shares this guidance because it treats AI as part of children’s lived environment and places responsibility where it belongs: with institutions that deploy systems at scale. The UNICEF Innocenti framework articulates clear expectations for how AI should be designed, governed, and overseen when it affects children’s rights, development, and opportunities. It provides a rights-based foundation that public institutions, educators, and civil society can use to ground real governance decisions.

We see this document as an important building block in a broader effort to treat AI as public infrastructure. It clarifies stakes, names responsibility, and reinforces the need for governance structures that endure across systems, providers, and political change.

Sharing this work supports our focus on governance that distributes power, protects those with limited agency, and embeds accountability into the systems that shape everyday life.