10-Minute Tool Cuts Deepfake Abuse at the Source

A simple 10-minute online tool is successfully changing how people think about deepfake technology, offering new hope in the fight against AI-generated image abuse.

Researchers at University College Cork developed "Deepfakes/Real Harms," a free educational intervention that significantly reduces users' intentions to engage with non-consensual synthetic intimate imagery. The tool has already been tested with over 2,000 people across different ages, genders, and digital literacy levels.

The research team discovered that people who engage with these harmful images often believe six common myths. Some think the images only cause harm if viewers believe they're real, while others assume public figures are fair targets for this kind of abuse.

The intervention tackles these myths head-on by encouraging reflection and empathy with victims. Results showed immediate impact, with effects lasting weeks after completion.

"Human users are the ones deciding to harass and defame people in this manner," said lead researcher John Twomey. "Our findings suggest that educating individuals about the harms of AI identity manipulation can help stop this problem at source."

The tool arrives as pressure mounts on platforms and lawmakers to address the rapid spread of AI undressing apps. Following the recent Grok AI controversy, the researchers argue that user education must be part of any comprehensive solution.

Dr. Gillian Murphy, the project's principal investigator, emphasized the importance of language when discussing this abuse. She noted that calling this material "deepfake pornography" is misleading since pornography implies consent, which victims never provide.

Why This Inspires

User feedback reveals the intervention's power comes from its non-judgmental approach. One participant shared that it "didn't come across as preachy" but instead offered "a pause button" to consider the human impact.

Another user appreciated how the tool addressed a blind spot in public discourse. "Too much of the deepfake discourse focuses on people being unable to tell them apart from reality when that's only part of the issue," they wrote.

The researchers built the project on their previous work in responsible software innovation. They propose a model where everyone from platforms to end users recognizes their power to minimize harm from emerging technologies.

Professor Barry O'Sullivan, who directs UCC's AI and Data Analytics program, called the work an urgent step toward improving AI literacy across society. He noted the dual benefit of reducing abuse while combating stigma faced by victims.

The intervention represents a refreshing shift in how we address technology-enabled harm by focusing on human behavior rather than just technical fixes or platform policy.