Surge in AI-generated deepfakes in South Korea election and sexual content misuse in Japan
Deepfake misuse has surged in Japan with sexual content targeting children and women posted on social media, while South Korea experienced a twenty-fold increase in election-related deepfakes compared to its previous election. Detection of high-quality deepfakes has become more difficult as they now incorporate realistic physiological signals.
What happened
In Japan, generative AI is being deliberately misused to create sexual deepfakes of real children and adult women, with fabricated nude images and videos being posted on social media platforms. In some cases, victims' names, home addresses, and school names are exposed alongside the fabricated sexual content.
In South Korea, deepfake videos have surged dramatically during the presidential election period, with deletion requests reaching nearly 10,000. The volume of deepfakes in the 2025 election has been reported as approximately 20 times greater than in the previous election.
Reported impact
Children are being victimized both as subjects of sexual abuse material and implicated as perpetrators when AI is used to generate such content. The volunteer monitoring group ひいらぎネット has documented large quantities of fabricated sexual images on social media, including school group photos where female students appear nude and yearbook photos with sexually modified content.
Deepfakes in South Korea's presidential election period have created significant challenges for platform operators and law enforcement in managing and responding to the synthetic media at scale.
Index analysis
Recent high-quality deepfake videos incorporate realistic physiological signals such as heart rate and subtle facial color changes that were previously thought to be undetectable in synthetic media, making detection increasingly difficult.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.