Japans Deepfake Crisis The Fight Against AI Pornography

A troubling trend is emerging in Japan as generative artificial intelligence is increasingly used to create deepfake pornography, causing significant harm to individuals from celebrities to schoolchildren. This new wave of digital exploitation is moving faster than the legal system can adapt, leaving victims vulnerable and authorities scrambling for solutions.

The Alarming Ease of Creation

The accessibility of this technology was highlighted by a recent case in which a 31-year-old company employee was arrested for creating and distributing sexual deepfakes of actresses. The man, who had no professional IT or AI training, told investigators he learned everything he needed from articles and videos online.

Using free image-generating AI software, he reportedly trained the system on images of 262 individuals, including actresses, pop idols, and announcers. He allegedly produced over 20,000 explicit images, which he sold through a subscription model on social media, earning approximately 1.2 million yen (S$10,200). His process was simple: take user requests, provide written instructions to the AI, and generate custom content for high-paying customers. This case demonstrates that creating convincing and harmful deepfakes no longer requires specialized expertise, only a computer and an internet connection.

A Widespread and Disturbing Trend

A 2023 survey by a U.S. security firm revealed the scale of the problem, finding that of nearly 96,000 deepfake videos online, a staggering 98 percent were pornographic. Japan is a significant part of this trend, with a flood of deepfake content targeting not only public figures but also underage girls.

The National Police Agency reported receiving over 100 complaints in 2024 concerning sexual deepfakes of minors. These images were often created using photos taken from school event pictures and graduation yearbooks, affecting students from elementary to high school.

The Human Cost of Digital Exploitation

The impact on victims is severe and long-lasting. Sumire Nagamori, a representative of the internet patrol group Hiiragi Net, emphasized the psychological damage. "The number of victims is increasing rapidly and some victims are unable to trust people because their names have been exposed," she stated. Nagamori believes the reported incidents are merely the "tip of the iceberg," as many victims are either unaware they have been targeted or are unable to speak out.

The Legal System Plays Catch-Up

The Japanese government has acknowledged the crisis and is discussing responses. A working group under the Children and Families Agency has recommended strengthening crackdowns and encouraging website administrators to remove the fake images. However, a comprehensive national plan is not expected until the end of the next fiscal year.

The core challenge is a legal one. Japan currently has no laws that directly regulate the production or distribution of deepfake pornography. In response, police are forced to use existing, often ill-fitting, laws. In some instances, charges for displaying obscene material have been successful. In others, authorities have tried to use defamation charges, arguing that placing a person's face on another body without permission damages their reputation.

However, this approach has limits. A senior police official noted there is no clear standard for applying defamation charges to an AI image generated entirely from scratch, even if it closely resembles a real person. Professor Takashi Nagase of Kanazawa University, a former judge, echoed this concern, stating, "Even if the image resembles a real person, it may not be considered defamatory if it is not the same person." This legal gray area leaves a significant loophole for perpetrators and a path of uncertainty for victims seeking justice.