In our study, one participant said they felt fearful the perpetrator would “continue uploading” images and the abuse might escalate to in-person violence. Others worried sexualised material associated with their names could be discovered by prospective employers or others.
This would require online services, including AI image-generation services, to take reasonable steps to keep users safe. It would also give the eSafety commissioner new powers to target nudify apps and websites.
Read more:
Australia set to ban ‘nudify’ apps. How will it work?
How are deepfakes made?
Technology companies must make their products and services safer and more resistant to misuse. This means building safeguards into AI tools to prevent abusive images being generated, making harmful content harder to find or circulate, and ensuring accessible reporting and rapid removal processes.
These findings suggest deepfakes can occur within peer cultures where sexualised content is used for humour, social bonding or mockery, as well as humiliation and harassment.
In 2017, a Reddit user, who went by the name “deepfakes”, shared non-consensual AI-generated videos of celebrities he’d created by “stitching” their faces onto the bodies of pornographic actors.
The burden should not fall on victims to continually search for and report the material.
The true prevalence is likely to be higher as many victim-survivors may never discover that images of them have been created or shared, while perpetrators may be reluctant to disclose their behaviour, even in an anonymous survey.
Four in five of those reports involve girls.
With little technical skill required, almost anyone can create these images and almost anyone can become a target.
What our research found
These tools are accessible, easy to find and sometimes free, expanding the pool of potential victims and perpetrators, including young people.
- people under 35
- LGBTQIA+ people
- Black, Indigenous, and other people of colour
- people with disabilities
- men.
These tools include “nudify” apps that digitally remove clothing from photographs, as well as AI generators that create new images or manipulate existing images from text prompts or reference images.
Some victim-survivors also described being disbelieved or having their experiences minimised, including by authorities, friends and family members because the images were considered to be “not real”.
Perpetrators must be held accountable. Prevention, however, cannot be left to individuals alone.
We also need consent-based and respectful relationships education which challenges the idea that creating sexualised images of others is harmless fun, and which works to address the gendered norms that normalise this behaviour.
What are the impacts?
As one young woman in the UK put it: “nobody cared, believed me or listened”.
A young non-binary person, also in the UK, said:
Educational interventions need to also focus on fostering social connection and building healthy relationships.
Most victim-survivors in our study reported anger, humiliation, shame, embarrassment, fear, helplessness and distress.
If this article has raised issues for you, call 1800RESPECT on 1800 737 732 or visit the eSafety Commissioner’s website for helpful online safety resources. You can also contact Lifeline crisis support on 13 11 14 or text 0477 13 11 14, Suicide Call Back Services on 1300 659 467, or Kids Helpline on 1800 55 1800 (for young people aged 5-25). If you or someone you know is in immediate danger, call the police on 000.
While men were more likely to report victimisation, women were more likely to report a broader range of harms and negative emotional and social impacts. Men were more likely to find the experience funny, although these were a minority.
Going further, Australia could consider following the UK in criminalising the development and supply of nudify tools.
Perpetrators described a range of motivations, not all of which involved revenge or an explicit intention to cause harm. Some “thought it was funny”, while others wanted to “experiment with technology” or “show off” to friends and online peers.
According to the Australian eSafety Commissioner, the number of reports from children involving digitally altered intimate images, including deepfakes, have more than doubled in the 18 months to June 2025 compared with the total number of reports received in the preceding seven years combined.
Prevention and accountability
Our team surveyed 7,230 adults in Australia, the UK and the US in July 2024. We found 7% reported being victims of non-consensual sexual deepfakes, with higher rates among:
Australia’s proposed digital duty of care, introduced into parliament last month, is an important step forwards.
Our recently published research – involving more than 7,000 people from Australia, the United Kingdom and the United States – reveals AI-generated image-based abuse is becoming increasingly widespread. It now affects ordinary people as well as public figures.
There should also be criminal offences for non-consensual sharing and creation of AI-generated intimate images in all states and territories.
We also found 3.2% admitted to having created, shared or threatened to share AI-generated intimate images of someone else without their consent.
Once images circulate, victims can face an exhausting “whack-a-mole” cycle of locating and reporting images, only for copies to reappear elsewhere.
The harms do not necessarily depend on how realistic the images appear to be. Victim-survivors can experience significant distress from losing control over how they are sexually represented. They can also experience fear others will believe the images are authentic, and the possibility of further humiliation or abuse.
There was kind of a level of scepticism from the people I told; because it wasn’t a ‘real image’ they couldn’t really see why I was so upset and thought it was less compromising because it wasn’t actually me in the picture.
Nearly ten years later, AI tools can generate highly realistic sexualised “deepfake” images in seconds, sometimes using just a single photograph. Often, high-profile people are targeted.
AI-generated image-based sexual abuse refers to the non-consensual creation, distribution, or threatened distribution of synthetic intimate images depicting real people. This is also known as “non-consensual sexual deepfakes”.
