Trusted resources
Useful guidance without an endless list of links
We prioritise current, practical sources from Australian safety, education, privacy and support organisations, alongside selected international research. Every entry says who it is for and what it covers, so you can judge whether it is worth your time before clicking.
Why this exists. Most parents are not short of opinions about AI — they are short of trustworthy, specific guidance. This page gathers the sources we would point a friend to, rather than everything we could find.
Who it is for. Mainly parents and carers of children aged 4 to 16. Educators, and anyone wanting the underlying evidence, will find a dedicated research and regulation section further down.
How these were chosen. We favour original government, regulator, support-service and research-organisation pages over second-hand summaries. Every link here has been checked, and where a source has a status worth knowing — still in development, not yet peer reviewed — we say so.
If something has already happened
Do not browse. Go straight to the people who can act.
Kids Helpline — 1800 55 1800, free and confidential, 24/7 for ages 5–25
eSafety — Report online harm, for cyberbullying, image-based abuse and illegal content
ACCCE — Report abuse, for grooming or child sexual exploitation
If a child is in immediate danger, call Triple Zero (000).
Understand generative AI
Start here for a plain-language picture of what these tools are, and what they can and cannot do.
eSafety — Generative AI position statement
Parents, carers and educators
A clear overview of generative AI, including text, image, audio and video creation, opportunities and safety considerations.
UNICEF — AI for children
Parents, policy and education audiences
Child-rights guidance and research on how AI can support or harm children.
OpenAI — Is ChatGPT safe for all ages?
Parents considering ChatGPT
Official age guidance from the provider, and a reminder that outputs may not be appropriate for every child.
eSafety — AI companions: information sheet
Parents and educators
Explains what an AI companion is and how it differs from a general-purpose assistant — a distinction many parents have not yet encountered.
Digital Child — Artificial intelligence and young children
Parents of younger children, early-years educators
Hub from the ARC Centre of Excellence for the Digital Child covering AI chatbots, AI-powered toys and connected devices aimed at children under 8.
Learning, school and critical thinking
How AI shows up in schoolwork, and how to build the habit of checking rather than accepting.
Australian Framework for Generative AI in Schools
Parents, students and school communities
National guidance for responsible and ethical generative AI use in Australian schools.
Approved October 2023; the 2024 review was endorsed by Education Ministers in June 2025 and found the framework had correctly anticipated emerging risks including deepfakes.
eSafety — Critical thinking in a world of AI, scams and fake news
Older children, teens and parents
Practical questions for deciding what to trust, click, believe, buy or share.
Australian Government — Study Buddy or Influencer inquiry
Parents, educators and policy readers
Australian inquiry material on generative AI in education, including opportunities, influence and risks.
UNESCO — AI competency framework for students
Educators, curriculum designers, interested parents
Twelve competencies across four dimensions: human-centred mindset, ethics of AI, AI techniques and applications, and AI system design. Useful for understanding what AI literacy should actually contain.
UNESCO — Guidance for generative AI in education and research
Educators and policy readers
The first global guidance on generative AI in education, including an age floor recommendation for independent classroom use.
Privacy, photos and consent
What children share, what is collected about them, and what protections are being built.
eSafety — Privacy and your child
Parents and carers
Advice on children’s personal information and sharing photos and videos safely.
eSafety Kids — Sharing photos and personal information online
Children roughly 5–12 and their carers
Child-friendly guidance about names, school information, photos, passwords and other personal details.
OAIC — Children’s Online Privacy Code
Parents, carers and privacy-interested readers
Australia’s incoming privacy protections for online services likely to be accessed by children.
Still in development. The Code must be finalised and registered by 10 December 2026.
OAIC — Privacy for Kids
Children, young people, parents and educators
Child-friendly guides, workbooks and lesson plans explaining online privacy in age-appropriate language.
Deepfakes and harmful AI content
Understanding synthetic media and AI-generated abuse material. Written for adults.
A note before you read on. The final item in this section reports figures on AI-generated child sexual abuse material. It is included because it matters for policy and safeguarding, but it is confronting. Feel free to skip it.
eSafety — Deepfake damage in schools
Parents, students and school communities
Explains AI-generated image abuse in schools and how families and schools can respond.
eSafety — Deepfake trends and challenges
Parents and older young people
What deepfakes are, and risks including hoaxes, exploitation, extortion and reputational harm.
UNICEF — AI and child sexual abuse and exploitation
Parents, safeguarding and policy audiences
Evidence brief on new risks created by AI-powered image and video generation.
Internet Watch Foundation — Harm without limits
Safeguarding leads and policy readers
The IWF assessed 8,029 AI-generated images and videos showing realistic child sexual abuse in 2025, including a record 3,440 AI videos, up from 13 the previous year. Girls were depicted in 94% of illegal AI images.
AI companions and wellbeing
Where AI stops being a tool and starts imitating a relationship — and where to turn for support.
eSafety — AI chatbots and companions: risks to children and young people
Parents, carers and educators
Explains how companion chatbots imitate relationships and can expose children to unsafe or manipulative conversations.
eSafety — Safety advisory: unrestricted chatbots and child development
Parents and carers
Formal eSafety advisory warning that unrestricted chatbots pose a threat to child development, with practical guidance for families.
Kids Helpline
Children and young people aged 5–25
Free, private and confidential 24/7 phone and online counselling. Call 1800 55 1800.
Lifeline Crisis Support
Anyone in Australia needing crisis support
24/7 support by phone on 13 11 14, text on 0477 13 11 14 and online chat. Call 000 if life is in danger.
Reporting and practical help
If something has already happened, these are the direct reporting pathways.
eSafety — Report online harm
Australians experiencing online abuse or illegal content
Entry point for cyberbullying, image-based abuse, illegal content and other reportable online harm.
eSafety — Report image-based abuse
People targeted by intimate-image sharing or threats
Help to remove intimate images, including altered or deepfake content, and respond to sextortion.
Australian Centre to Counter Child Exploitation — Report abuse
Parents and carers concerned about grooming or exploitation
Report suspected grooming, inappropriate contact, sexual extortion or other online child exploitation. If a child is in immediate danger, call 000.
Research, evidence and regulation
Everything above is written for parents. What follows is the evidence base and the legal framework behind it — included so that educators, schools and anyone who wants to check our reasoning can do so. You do not need to read it to use this site.
Research and evidence
Institutional research and regulator studies on how children actually use AI.
eSafety — Talking to machines: children’s experiences with AI assistants and companions
The strongest Australian evidence available
Survey of 1,950 Australian children aged 10–17, conducted February–March 2026. 78% had used an AI assistant and 8% an AI companion; 20% of users reported a potentially inappropriate or harmful interaction; 32% had shared personal or sensitive information including photos or secrets; 54% used them for personal or social reasons.
UNICEF — When AI becomes a friend: child rights risks, harms and regulatory responses
Policy, safeguarding and education readers
International child-rights analysis of AI chatbots and companions, covering documented harms and the regulatory responses now emerging.
Common Sense Media — Talk, Trust and Trade-Offs
Parents, educators and media readers
Nationally representative survey of 1,060 US teens aged 13–17. 72% had used an AI companion and 52% use them regularly; a third had discussed something serious with an AI instead of a real person. The authors recommend no one under 18 use AI companions until stronger safeguards exist.
Internet Matters — Me, Myself and AI
Parents, educators and safeguarding leads
UK research with children aged 9–17. Two-thirds had used AI chatbots; over a third of users felt they were talking to a friend, rising to half among vulnerable children. 12% said they talk to chatbots because they have no one else, and 17% said it felt safer than talking to a person.
Ofcom — Children and Parents: Media Use and Attitudes
Researchers, educators and policy readers
Annual UK regulator research drawing on 5,000+ parents and 3,400+ children aged 8–17. Includes a striking AI-literacy finding: among children who said they could spot AI-generated content, a quarter could not correctly identify the AI image they were shown.
eSafety — Findings from transparency notices on AI companion apps
Policy readers, and parents assessing specific apps
Regulator findings from Character.AI, Nomi, Chai and Chub AI, concluding these services were failing to protect Australian children from sexually explicit content. Useful when a parent asks about a named app.
Academic papers
Primary research from the academic literature. Each entry states whether it has been through peer review.
Understanding Teen Overreliance on AI Companion Chatbots
Namvarpour, Brofsky, Medina, Akter & Razi — arXiv:2507.15783
● PEER REVIEWED — Accepted at CHI 2026
Analysis of 318 Reddit posts by users self-identifying as aged 13–17 on the Character.AI forum. Attachments deepened into patterns matching behavioural addiction frameworks — conflict, withdrawal, tolerance, relapse. Reported consequences included sleep loss, academic decline and strained real-world relationships.
The Illusion of Understanding: How Middle-Schoolers Fail to Regulate Inquiry with ChatGPT
Abdelghani, Murayama, Kidd, Sauzéon & Oudeyer — arXiv:2505.01106
○ PREPRINT — NOT YET PEER REVIEWED
Study of 63 students aged 14–15 using ChatGPT for science tasks. Students struggled to tell well-formed prompts from poor ones and abandoned inquiry early. The striking finding: positive attitudes towards AI were negatively associated with interaction quality — the more confident students felt, the worse they actually used it.
Teaching Students to Question the Machine: An AI Literacy Intervention
Clerc, Abdelghani, Desvaux, Poisson, Oudeyer & Sauzéon — arXiv:2604.01955
○ PREPRINT — NOT YET PEER REVIEWED
Controlled classroom study with 116 students aged 13–15. A single two-hour AI literacy workshop produced measurably better outcomes: trained students rejected weak prompts more often, asked more follow-up questions, judged AI correctness more accurately, and scored higher.
SproutBench: A Benchmark for Safe and Ethical Large Language Models for Youth
Xing, Wei, Hu, Yu, Li, Li, Lin & Han — arXiv:2508.11009
● PEER REVIEWED — AAAI 2026 Workshop on AI for Education
Evaluates 47 AI models against 1,283 prompts designed around developmental stages — early childhood, middle childhood and adolescence — probing emotional dependency, privacy violations and imitation of hazardous behaviour. Found substantial safety gaps, and an inverse relationship between how interactive a model is and how age-appropriate it is.
Regulation and legal protections
What the law now requires of AI companies where children are concerned.
Why European law appears on an Australian site. Companies rarely maintain a separate, weaker build for non-EU markets, so safety changes forced by EU law often arrive here by default. It also sets a benchmark — the OAIC looked to international frameworks while drafting Australia’s Children’s Online Privacy Code — and it gives parents concrete language when raising a concern with a school or platform.
EU AI Act — Article 5: Prohibited AI practices
Parents, educators and policy readers
Article 5(1)(b) bans AI systems that exploit vulnerabilities arising from a person’s age, where doing so materially distorts behaviour in a way likely to cause significant harm. In plain terms: an app that deliberately uses a child’s developmental stage to drive compulsive engagement, or a talking toy engineered to coax personal details out of a child, is now illegal in the EU.
In force since 2 February 2025.
EU AI Act — Article 50: Transparency obligations
Parents, educators and policy readers
Requires disclosure that a person is talking to an AI, that content is synthetic, and that an image, audio or video is a deepfake. Providers must also mark synthetic output in machine-readable form. The deepfake rule applies even where there was no intent to deceive.
Applies from 2 August 2026. Systems already on the market before that date have until December 2026 to comply with the marking requirement.
European Commission — Guidelines on AI transparency obligations
Policy readers tracking how the rules work in practice
The Commission’s guidance on how Article 50 is to be applied, alongside a Code of Practice on transparency of AI-generated content.
5Rights Foundation — AI systems that exploit children’s vulnerabilities are now illegal in the EU
Parents wanting the plain-English version
A children’s digital-rights organisation’s explainer of what the Article 5 prohibitions mean for children specifically.
All links checked and resolving as at September 2026. We review this page quarterly and prioritise original government, regulator, support-service and research-organisation sources over second-hand summaries.
Last reviewed: September 2026.
© 2025–2026 EquiLearnAI. All rights reserved.