Online Safety and Children's Digital Lives: Report of a National Survey of Children and their Parents/Caregivers

Key Findings

• — 43% of children aged 9 to 17 used generative AI in the month before interview; use was 16% among 9-to-11-year-olds and 67% among 16-to-17-year-olds, and the report names age as the strongest factor.
• — 3% of all children used generative AI to talk about their worries and get advice (1% at 9-11, 3% at 12-13, 7% at 14-15, 6% at 16-17; 5% of girls against 2% of boys; 5% of children with a disability), and 6% used it for advice on physical health or fitness (13% at ages 14 to 17).
• — The dominant uses were schoolwork: 21% wrote essays or stories and 20% summarised or explained longer texts; 12% sought recommendations; 6% created images or video; 1% reported making deepfakes.
• — In the past year 21% of children were exposed to self-harm content, 21% to content on ways to be very thin, 24% to drug-taking content, 23% to violent content, 23% to dangerous challenges, and 13% each to suicide content and pornography.
• — Parents were often unaware of what their child had seen: the report contrasts children's reported exposure with parental awareness across content types, including suicide and eating-disorder content.
• — The regulator frames the study against the growth of generative AI chatbots and assistants and announces a public information campaign on artificial intelligence for 2027.

Methodology Notes

Nationally representative in-home face-to-face survey by Ipsos B&A (Ipsos Ireland), n = 1,002 children aged 9 to 17 (weighted totals reported per subgroup) plus one parent or caregiver per child, fieldwork July 2025 to January 2026, self-report with a past-month recall window on the generative-AI items (Table 5, page 47) and a past-year window on harmful-content exposure. The PDF's creation date is 2026-09-07 and the regulator's announcement page carries a 2026-09-10 publication time. Generative AI is one short section of a broad online-safety survey; the items ask about purpose of use, not about which products were used. Ireland also fielded the EU Kids Online 2026 comparative survey, which reports 14% of children across 19 countries talking to a chatbot about worries, so the two Irish estimates come from different instruments and samples.

Sources