About one in 14 adults in Australia, the United Kingdom and the United States reported that someone had created, shared or threatened to share sexually explicit images generated by artificial intelligence, according to a peer-reviewed study of 7,230 people. The research found that this form of image-based sexual abuse was associated with substantial emotional and personal harm, particularly among women and members of several marginalized groups.
“The findings highlight the extent of abuse across the wider community,” said study author Professor Nicola Henry, a researcher at RMIT University.
“We now have evidence of how far-reaching abuse based on AI-generated images is impacting not just celebrities and public figures, but people across communities.”
“The actual scale may be even higher, as some people may not even know that an AI-generated image of them was created or shared.”
Rates of AI-generated image abuse
Henry and her colleagues surveyed 7,230 adults in Australia, the United Kingdom and the United States in July and August 2024.
Overall, 15.1% of respondents reported at least one experience of abuse involving digitally altered intimate images, including both AI-generated fakes and images created using other editing methods. About 7%, or roughly one in 14 respondents, said artificial intelligence was specifically involved.
Among all respondents, 4.3% said someone had created an intimate AI-generated image of them without their consent, 4.4% said such an image had been shared, and 3.6% said they had been threatened with its distribution. Threats to share intimate images are commonly used in sextortion schemes.
The United Kingdom had the highest reported rate of AI-related abuse, at 9.3%, followed by Australia at 7.7% and the United States at 6.5%.
Victimization rates were higher among people under 35, LGBTQ+ respondents, people of color and people with disabilities.
Emotional and personal impact
Among respondents who had been victimized, 90.3% reported at least one negative emotion and 86.7% reported at least one negative impact on their lives.
Women were more likely than men to report exclusively negative reactions, at 87.6% compared with 71.2%. They were also more likely to fear for their safety, at 42.4% compared with 18.1%, to report feeling sad or depressed, and to describe effects on their mental health.
Men were significantly more likely than women to report neutral or positive reactions, such as finding the images funny. However, those reactions were rare overall.
Who was responsible?
Most victims, 55.6%, said the perpetrator was someone they knew only online. Nearly 30% said they had interacted directly with the person, while 14.6% were unsure who was responsible.
When asked about their relationship with the perpetrator, the most common response was “strangers,” at 21.4%, followed by friends at 14.2% and ex-partners at 9.2%.
Perpetrators were most often reported to be male, at 47.4%, while 36.7% of victims were female. The authors caution that this result may partly reflect scammers posing as women online.
“As generative AI becomes more widely available, there is an increased likelihood that these tools will be misused to create, share, or threaten to share intimate images without consent,” Henry said.
“What we don’t know yet is how quickly that translates into increased damage.”
The findings were published online on September 8, 2026, in the International Journal of Human-Computer Interaction.
Nicola Henry et al. “That experience still haunts me”: A multi-regional study of sexual abuse victimization based on AI-generated images. International Journal of Human-Computer Interaction, published online September 8, 2026. doi: 10.1080/10447318.2026.2723677
Source: www.sci.news