A photograph shared online could become the source of months of anxiety, humiliation and emotional distress after artificial intelligence (AI) transforms it into a sexually explicit image without its owner's consent.

Unlike traditional image-based sexual abuse, AI-generated deepfakes do not require someone to possess genuine intimate photographs. A person's face can be digitally placed onto fabricated nude images or sexual scenarios, creating convincing material that can be circulated without their knowledge.

Now, new research suggests the consequences extend far beyond embarrassment, with victims reporting psychological distress, damaged relationships and disruption to their professional lives.

A study published in the International Journal of Human-Computer Interaction found that approximately one in 15 adults (7%) across the United States, United Kingdom and Australia had experienced at least one form of AI-generated image-based abuse since turning 18.

Of those affected, 87% reported consequences involving their health, wellbeing, relationships, careers or finances.

How AI-Generated Sexual Images Affect Mental Health

Research published in the journal Women's Health on October 23 examined the experiences of 25 Australian adults, including 15 survivors of sexualised deepfake abuse and 10 perpetrators.

Participants described anxiety, depression, shame, diminished self-worth and social withdrawal. Some also reported difficulties maintaining relationships and continuing their education or employment.

One particularly distressing aspect was the uncertainty surrounding who had seen the fabricated images and whether others believed they were authentic.

Researchers found that survivors frequently felt compelled to explain or disprove images depicting them in sexual situations they had never participated in.

This created an additional psychological burden, with some participants describing a loss of control over their identity and reputation.

The researchers concluded that the fabricated nature of deepfakes introduces distinctive challenges beyond those associated with conventional image-based sexual abuse.

Read more

Why Fake Images Can Cause Real Psychological Trauma

Although the images do not represent genuine sexual encounters, the humiliation, fear and violation experienced by victims are real.

The possibility that fabricated material could resurface, spread to new audiences or damage existing relationships can create prolonged uncertainty.

This can contribute to persistent anxiety, hypervigilance and avoidance of social situations. Some survivors may also experience symptoms associated with depression and trauma-related distress.

The research highlights how the digital nature of the abuse can prolong its consequences. Unlike an isolated incident, an image uploaded online can be copied, redistributed and encountered repeatedly.

A survey also found differences in how victims responded according to gender. Women were more likely than men to report depression (43% versus 26%) and fear for their personal safety (42% versus 18%).

However, the researchers cautioned that experiences of abuse were not confined to women. Men, younger adults, ethnic minorities and people with disabilities also reported elevated victimisation rates.

How AI Tools Are Making the Problem Worse

The growing accessibility of AI image-generation applications has lowered the technical barriers to creating sexually explicit deepfakes.

Research from the University of Florida examined the online communities and resources supporting the creation and distribution of synthetic non-consensual intimate imagery.

Researchers analysed 360,462 posts and submissions across 4chan and Reddit, identifying networks through which users exchanged tools, tutorials and techniques.

The findings highlighted how online communities can help inexperienced users develop the ability to produce and distribute abusive material.

What Can Victims Do After Deepfake Abuse?

Experts say tackling the problem requires more than asking victims to report abusive images after they appear online.

Researchers have called for stronger platform safeguards, improved detection technologies and measures targeting the tools and online communities enabling the abuse.

For people affected, preserving evidence, documenting threats and reporting abusive content to the relevant platforms can help support removal efforts and potential investigations.

Seeking support from trusted people and qualified mental health professionals may also help victims manage anxiety, emotional distress and the longer-term consequences.

Crucially, victims should not blame themselves for images created or distributed without their consent.

The central finding is difficult to ignore: an image does not have to be real to cause lasting harm. As AI tools become increasingly accessible, protecting people's digital identities is also becoming an important part of safeguarding their psychological well-being.

Read more