Confronting accounts have given a rare insight into the minds of perpetrators of deepfake image-based abuse.
People who create fake and often sexualised images of others do so largely out of curiosity and a desire for revenge, according to a new report examining their motivations.
As part of a Monash University study, researchers interviewed 10 perpetrators of deepfake image abuse and 15 victim survivors.
It found that perpetrators frequently downplayed the behaviour, which overwhelmingly targeted women, and that it was normalised amongst young men.
Perpetrator interviews suggested most offending was motivated by money, or sextortion, curiosity, wanting to cause harm or seek revenge and peer reinforcement.
'Slippery slope' to offending
Lead author of the study and professor of criminology at Monash University, Asher Flynn, said the first-of-its-kind report provided invaluable insight into offending behaviour.
"While a number of the perpetrators were engaging in these behaviours to harm or to get back at someone, others were also doing it because they thought it was funny or because they were able to access the [Artificial Intelligence] tools, she said.
"[They] sort of went down this slippery slope of starting to investigate what might be possible and then actually getting into really harmful behaviours and sharing these with their friends, getting reinforcement from them."
One participant described how he altered sexualised deepfakes of a colleague after a relationship breakdown, sharing the images to a work-related WhatsApp group and another that included the victim's daughter.
"I'm normally not a person who would hurt anybody… and I was just really surprised in retrospect," he told the study.
"All the evil thoughts I had to get back at this person, because normally I would not behave in that way.
"It was a reaction for what she did and that [the break-up] hurt significantly."
'A little bit of revenge'
Another participant told the study they were merely curious to try the deepfake technology.
"You have a little god-like buzz of seeing that you're capable of creating something like that," they said.
Another man interviewed for the study said he created a fake profile on three dating apps using photos from his ex's social media.
He told researchers he used AI software to alter them into sexualised deepfake images.
"He broke up with me, and he cheated on me… he deserved this," he told the study.
"I was very angry at that time. I just wanted a little bit of revenge."
Professor Flynn said perpetrators experienced a disconnect between their behaviour and the harm it was causing to victims.
"It's pretty challenging when you're hearing people reflect on their behaviours, particularly when they are minimising the harms that they've caused or looking at ways to justify the behaviours they've engaged in," Professor Flynn said.
Victim blaming concerns
Deepfake victim-turned lawyer Noelle Martin's fight against image-based abuse began as a teenager when she discovered photos from her social media had been doctored into fake intimate images and distributed online.
"It was shocking because at the time, I didn't actually know what the term was to even describe this," she said.
"It wasn't dominating the news like it is today and I didn't know what was happening, why it was happening, who was behind it."
The altered images were reported to police, but the perpetrators have never been held to account.
Victim survivors interviewed as part of the Monash University study also reported little to no recourse for perpetrators.
Professor Flynn said the trauma for victims was compounded by knowing the fake images existed online and the prevalence of victim-blaming behaviour.
"This victim-blaming attitude or a mentality around that women's bodies are just there for the use and access of other people… was a really problematic issue that came out in the study," she said.
Enforcement push
The federal government last year criminalised sharing non-consensual sexually explicit deepfakes, with perpetrators facing penalties of up to seven years in jail.
Meanwhile, in September, the NSW government passed legislation outlawing the use of AI to create intimate images of a real, identifiable person without their consent.
Ms Martin has dedicated her legal career to campaigning for such laws but said she believes enforcement of the legislation was still lacking.
"My personal experiences have motivated me to fight this because I don't want people to go through what I went through," she said.
"The gaps in Australia mainly relate to the fact that while we might have laws… we are dealing with a massive enforcement issue."
Children at risk
Last month, the eSafety Commission also blocked Australians from access to three major "nudify" services linked to the generation of child sexual exploitation material.
NSW Women's Safety Commissioner Hannah Tonkin said there was a "disturbing increase" of the issue, particularly amongst school-age children.
"When I speak to women and to girls, including school students, they are terrified that they are going to be the subject of a deepfake image," Ms Tonkin said.
"This technology dehumanises women, it degrades women, it humiliates women and it has really long-lasting impacts.
"If a woman or a girl is the target of an AI-generated deepfake image, they should report that to police because this is a crime."
