Skip to main content
NGO report Credible — Major labs, established NGOs, reputable named-author preprints

AI- and non-AI online child sexual victimisation, disclosures and help-seeking: A national study of Australian adolescents

Nationally representative online panel survey of 1,894 Australians aged 16 to 18, fielded March to April 2026, asking about online sexual victimisation before age 18 (non-consensual sharing of sexual images and online sexual solicitation by an adult, measured with the Australian Child Maltreatment Study adaptation of the Juvenile Victimization Questionnaire), whether artificial intelligence was involved, to whom victims disclosed (parents, family, friends, authorities, or a generative-AI platform such as ChatGPT or Claude), peer-group experiences, help-seeking from AI, and worry about AI-generated fake images. It is the first Australian population estimate of AI involvement in online child sexual victimisation and reports that victims disclosed to an AI more often than to any authority.

Publisher

Adelaide University, supported by ICMEC Australia (International Centre for Missing & Exploited Children Australia)

Published

6 Jul 2026

Added

today

Key Findings

  • Before age 18, 6.4% had a sexual image shared without consent and 11.3% were sexually solicited online by an adult; girls reported higher rates of both (+2.5 and +6.8 percentage points)
  • AI was involved in 26.0% of non-consensual image-sharing cases and 8.6% of adult-solicitation cases (14.9% of online sexual victimisation overall); boys were more likely than girls to report AI involvement (37.5% vs 19.7%)
  • 65.7% of victims disclosed to someone; 18.7% disclosed to an AI against 13.2% to any authority (teacher, doctor, counsellor, police, helpline or reporting portal); boys disclosed to AI at 27.1% vs girls 14.6%; among those who disclosed to an AI only 18.9% also told a human, and 35.4% of victims told no one or only an AI
  • Lifetime AI use under 18 was 80.6% (56.1% chatting with an AI, 45.1% generating images, 6.5% AI embedded in social platforms, 1.7% generating pornographic content); 36.6% had asked an AI for personal life advice; 19.0% of boys and 15.4% of girls used AI for help-seeking at least weekly
  • 22.8% knew a friend who had experienced online sexual victimisation (28.6% personal or peer experience combined); 33.9% were at least moderately worried about AI being used to create fake images of them (girls 38.5%, boys 29.7%)

Methodology Notes

Cross-sectional online panel survey with proportional quota sampling and post-stratification weighting by age, sex and location; final analytic sample n=1,894 aged 16 to 18 reporting on experiences before age 18; fieldwork March to April 2026 (month precision); victimisation items adapted from the Juvenile Victimization Questionnaire R2 (ACMS version); disclosure to AI captured separately from AI-enabled reporting portals; 95% confidence intervals reported throughout; limitations stated (retrospective self-report; gender-diverse subsample too small to disaggregate). Report page dated 6 July 2026; PDF 58 pp, created 5 July 2026, uploaded August 2026; DOI 10.25909/32902481; companion papers: 'Prevalence and disclosure of AI and non-AI online child sexual victimisation' (SSRN 10.2139/ssrn.7000199, full data tables) and 'The digital lives of adolescents' (10.25909/32904020).

Authors

Timothy I. C. Cubitt (Adelaide University), Katie Logos (Adelaide University), Russell Brewer (Adelaide University), Ben Mathews (Queensland University of Technology)

Tags

australianciideepfakesdisclosurehelp-seekingadolescentsicmecnational-prevalence

Cite This

APA

Timothy I. C. Cubitt (Adelaide University) et al. (2026). AI- and non-AI online child sexual victimisation, disclosures and help-seeking: A national study of Australian adolescents. Adelaide University, supported by ICMEC Australia (International Centre for Missing & Exploited Children Australia). https://icmec.org.au/wp-content/uploads/2026/08/Cubitt_et_al_AI_and_non_AI_online_child-sexual_victimisation_disclosures_and_help_seeking-1.pdf