AI-Generated Sexual Deepfakes of Celebrities and Students Arrested in Japan
Police in Japan arrested a 31-year-old man who used freely available image generation AI to create over 500,000 sexual deepfake images of approximately 300 female celebrities and idols and posted them on paid membership websites, earning about 11 million yen. Separately, sexual deepfake images of female students were created and shared at an Okinawa middle school, causing significant distress to the victims.
What happened
A 31-year-old man from Sapporo was arrested on January 19, 2026 on suspicion of displaying obscene digital records after creating and publishing sexual deepfake images of female celebrities and idols using freely available image generation AI on membership-based websites.
Between December 2024 and May 2025, the man allegedly uploaded tens of sexual deepfake images to two membership websites, made them accessible to the general public, and earned approximately 11 million yen in viewing fees. The man used publicly available image generation AI trained on photographs of actresses and idols to create the images, then promoted samples on X and Instagram to recruit subscribers to his paid sites.
Police estimate the man created and posted over 500,000 sexual deepfake images of approximately 300 female entertainers and idols on the sites since summer 2023. The subscription model allowed high-tier paying users to request specific celebrities to be featured in new images.
Reported impact
At an Okinawa middle school on the main island, sexual deepfake images of female students were circulated after being created and shared among acquaintances, though large-scale online dissemination was prevented. The affected students experienced significant shock and suffering from the incident.
Specialists note that AI services with the capability to digitally undress people in photographs with a single click have been identified online, and sexual deepfakes derived from school event photographs and yearbook images have become widespread problems in Japan.
Index analysis
Security company Trend Micro highlighted that no law currently directly prohibits the creation of sexual deepfakes, leaving a regulatory gap in legal protections. An expert at the company stated that AI service users should recognize that such images can damage victims' dignity and that the regulatory framework requires reconsideration.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.