Deepfakes: What Australian Parents Need to Know

Quick answer

A deepfake is AI-generated audio, image or video showing a real person doing or saying something they never did. The tools that make them are now free, fast and simple enough for a bored Year 9 student to use, which is why the harms have moved from celebrity gossip and election coverage into school group chats. For parents, the job has three parts: understand what deepfakes are, talk about them with your teenager before an incident rather than after, and know the reporting path if something goes wrong - the school for school matters, the eSafety Commissioner for image-based abuse. None of this requires technical skill. It requires the steady parenting you already do, pointed at a newer problem.

What a deepfake actually is

The word joins "deep learning" - the technique behind modern AI - with "fake". A deepfake is made by feeding an AI system real photos, footage or voice recordings of a person, then generating new material that looks or sounds convincingly like them. A few years ago this needed serious computing power and real skill. Now it takes a phone, an app and a handful of images lifted from a public social media profile.

Three forms matter most for families:

One uncomfortable truth up front: you cannot reliably pick a good deepfake by eye, and neither can your teenager. The old tells - strange hands, odd blinking, warped backgrounds - date quickly as the tools improve. The protection is not sharper eyes. It is better habits, which we will get to.

Why this belongs on a parent's radar

Deepfake harm in a school context is rarely about politics. It is personal. Incidents involving fabricated images of students and teachers have already reached Australian schools, and the pattern is consistent: material gets made as a "joke" or a weapon, spreads through group chats faster than any adult hears about it, and lands hardest on the person depicted.

Three things make this worth your attention now. First, the harm is real even when everyone knows the material is fake - humiliation does not require authenticity. Second, the sexualised category is the most serious: fabricated intimate images of a real person are image-based abuse, and when the person is under 18 the stakes rise sharply. Third, your child could be on either side. A teenager who makes or shares a deepfake of a classmate is not pulling a prank, and sharing sexually explicit deepfakes of a real person without consent can be a criminal offence in Australia.

There is also a quieter cost: a generation tempted to conclude that nothing on a screen can be trusted. That cynicism is worth heading off as deliberately as the abuse - part of the wider picture in whether AI is safe for teenagers.

How to talk about it before anything happens

Open with curiosity, not a lecture: "Have you seen deepfakes going around? What's the worst one you've heard about?" Most teenagers have seen more of this than their parents and will happily explain it, which usefully puts them in the expert seat and you in the listener's.

Then land two messages, separately and calmly.

If it ever happens to you or a friend: it is not your fault, you will not be in trouble, and telling me early makes everything easier to fix. Do not forward the material to anyone, even to show how bad it is - forwarding spreads the harm.

If you are ever tempted to make one: a fabricated image of a real person is not a joke, whatever the group chat says. It is abuse of a real person, schools treat it that way, and so can the law.

Finally, shrink the raw material. Deepfakes are built from public photos, videos and voice clips, so what your teenager shares publicly matters - a conversation covered properly in what teenagers share online and why it matters.

If your child is targeted: the practical steps

1. Steady them first. Say the important sentence out loud: this is not your fault.
2. Capture evidence - screenshots, usernames, URLs, dates - without resharing anything.
3. Report it in the app where it is circulating; most platforms have a route for fake or intimate imagery.
4. Involve the school if other students are part of it. Ask what happens next and when.
5. Report to eSafety. The eSafety Commissioner runs reporting schemes for image-based abuse and for serious cyberbullying of children, and can require platforms to remove material.
6. Contact police if the material is sexualised and your child is under 18.
7. Watch the aftermath - sleep, mood, reluctance to go to school - and bring in your GP or the school counsellor if the distress lingers.

The longer game: raise a verifier

Since detection by eye is a losing game, teach provenance instead - where a piece of media came from. The questions that work are simple: who posted this first, is any outlet you trust carrying it, and does the person supposedly involved confirm it? A claim that exists only in one screenshot is a claim, not a fact.

This is the same verification muscle that protects teenagers from AI's confident errors in schoolwork, built step by step in teaching teenagers to fact-check AI. Deepfake resilience is not a separate skill; it is AI literacy applied to media, one strand of the broader project described in AI education for teenagers in Australia.

The recommendation: have the conversation this week, while it is still hypothetical. Cover both sides - being targeted and being tempted - agree that nothing ever gets forwarded, and make sure your teenager knows the eSafety Commissioner exists and that coming to you early is safe. A calm ten-minute chat now beats a crisis response later.