Field: Technology

AI-Generated Sexual Imagery Abuse is Far More Common than Many Assume

Published | Technical Staff

AI-Generated Sexual Imagery Abuse is Far More Common than Many Assume

Visualization

In an age where artificial intelligence is rapidly proliferating across various sectors, a troubling correlation has emerged highlighting the dark side of digital innovation: the prevalence of AI-generated sexual imagery abuse. A peer-reviewed study led by Professor Nicola Henry from RMIT University reveals that about one in 14 adults across Australia, the United Kingdom, and the United States has encountered the harrowing experience of someone creating, sharing, or threatening to disseminate sexually explicit imagery of them generated through AI.

The study involved an extensive survey of 7,230 respondents conducted in July and August of 2024, aimed at quantifying the incidence and impact of this modern form of abuse that transcends the traditional narratives of victimization which typically center on public figures and celebrities. This research offers critical insights into a burgeoning societal concern, suggesting that the ramifications extend far beyond a limited cohort, and instead, affect a wide swath of the population.

The findings are stark: a striking 15.1% of participants reported experiencing abuse related to digitally altered intimate images, with AI playing a direct role in roughly 7% of those accounts. Specifically, 4.3% reported non-consensual creation of AI-generated intimate images, while 4.4% indicated that such images were shared without their consent, and 3.6% disclosed threats of sharing, a tactic often associated with sextortion schemes. The geographic breakdown of these figures highlights the issue's geographic variance, with Britain reporting the highest incidence (9.3%), followed closely by Australia (7.7%) and the United States (6.5%).

Younger individuals, particularly those under 35, and marginalized groups such as LGBTQ+ persons, people of color, and those with disabilities, were disproportionately affected. Among the victims, an overwhelming 90.3% experienced at least one form of negative emotional response, while 86.7% reported adverse impacts on various facets of their lives. Gender differences in reactions became evident, with women demonstrating a higher likelihood to report exclusively negative emotions—87.6% compared to 71.2% for men—and experiencing heightened feelings of fear for their safety (42.4%) along with other negative mental health outcomes.

Interestingly, the data also revealed a divergence in how men and women respond to these encounters, with men more frequently describing their reactions as neutral or even finding humor in the incidents, albeit these reactions were rare overall. The relationships between victims and perpetrators were notably diverse, with 55.6% identifying the perpetrator as someone known solely through online interactions, while nearly 30% had engaged with their abuser in real-life scenarios. The most common descriptor for the relationship was 'stranger' (21.4%), with a smaller proportion indicating friendships or previous romantic involvements.

The authors caution against oversimplified views of perpetrators; while 47.4% of victims noted that the abuser was male, a significant 36.7% identified a female perpetrator, possibly reflecting cases where scammers masquerade as women online, complicating the landscape of accountability and victim support.

As generative AI tools become increasingly accessible and sophisticated, concerns about their potential misuse grow more pronounced. Professor Henry articulates the urgency of understanding how the widening availability of these technologies may catalyze an increase in victimization rates, stating, "What we don’t yet know is how quickly that is translating into increased victimization." Her statement underscores the pressing need for further research to quantify this escalating risk and formulate effective strategies for prevention and support.

Published in the International Journal of Human-Computer Interaction on September 8, 2026, these findings add a vital dimension to ongoing discussions regarding digital ethics and personal safety in an increasingly AI-driven world. As society grapples with the complexities of technological advances, the conversation surrounding AI-generated sexual imagery is likely to intensify, illuminating the challenges we face in safeguarding individual rights in the digital age.