Source: Policy Options (IRPP)

Children’s privacy should be treated as critical public infrastructure in the AI era, because AI systems now shape what children see, how they learn, how they are classified, the online risks they face, and how institutions interact with them. When those systems collect, infer, retain, or expose too much about children, harms extend beyond a single app and undermine trust across education, public services and digital life. The piece cites the OECD’s 2025 report and UNICEF’s 2025 guidance to show that digital services bring both opportunities and significant risks that require safety, transparency, fairness and explicit protection of children’s data. It also notes the Office of the Privacy Commissioner’s 2026–27 departmental plan and the 2026 Global Privacy Enforcement Network sweep as evidence that regulators are focusing on AI implications for children’s privacy.

The author urges the federal government to act in four interconnected areas: require any child-facing AI to undergo a child-specific privacy-and-rights impact assessment before deployment; exclude children’s data by default from model training, secondary use and speculative analytics unless a strong public-interest case and safeguards exist; require retention limits, audit logs, plain-language notices, and simple deletion or human-review paths in public-sector AI contracts; and mandate disclosure when youth-facing AI shapes an interaction, recommendation, risk score or decision. The argument is reinforced by concerns about synthetic harms — including AI-generated image and video tools that escalate child sexual-abuse risks — and by the OECD’s 2025 assessment that age-assurance alone is insufficient, supporting a child-centred governance frame combining oversight, safety, privacy, fairness, accountability and children’s best interests.