Japans Alarming Rise Of AI Sexual Deepfakes

Generative AI is fueling a rapid spread of sexual deepfake images and videos in Japan. This alarming trend primarily targets and harms children and women.

Deepfake images, partially modified, are seen on a computer in this April 24, 2025, file photo. (Mainichi/Yuki Machino)

The Escalating Threat of AI Generated Abuse

In Japan, a disturbing trend is rapidly gaining momentum: the creation and dissemination of sexual deepfake images and videos. These fabricated materials, powered by generative artificial intelligence, are increasingly targeting vulnerable populations, particularly children and women. The Mainichi Shimbun has investigated the tangible harm caused by this misuse of AI, which not only victimizes young people but also has the potential to turn unsuspecting individuals into perpetrators through the ease of content creation.

Instances abound where fake nude images of individuals are crafted without their consent. These deeply personal violations are then often exposed online, frequently accompanied by the victim's real name, address, and even school information, leading to devastating consequences.

How Deepfakes Are Created and Spread

Sumire Nagamori, who heads the volunteer organization Hiiragi Net, dedicates her efforts to patrolling the internet. Her organization reports harmful content to platform operators and the police. With a grim expression, she described sifting through an overwhelming volume of sexually explicit fake images and videos proliferating on social media. Examples include school group photos maliciously edited to depict female students as naked, and yearbook pictures where the bodies beneath the faces have been sexually altered.

A significant concern is the exploitation of school photo systems. Many schools engage professional photographers who upload event photos to dedicated websites, enabling families to select and purchase images. It appears these private photo collections are being leaked and misused. Disturbingly, online forums even host exchanges like, "Let's trade login IDs and passwords for each school's photo sales website."

Generative AI Lowering the Barrier to Harm

Previously, creating sexually explicit fake images, such as "idol collages" (known as "aikora") that merge a person’s face with pornographic material, demanded a certain level of editing knowledge and technical skill. However, the landscape has dramatically changed with the advent of apps and websites incorporating generative AI. Users can now simply upload an image and generate a sophisticated deepfake in less than a minute. This significantly lowered barrier to creation is a primary factor contributing to the rapid escalation in these harmful activities.

The Devastating Impact on Victims and Avenues for Support

The Tokyo based nonprofit Organization for Pornography and Sexual Exploitation Survivors, or PAPS, has been a crucial support system for victims of digital sexual violence. Around six months ago, PAPS began actively encouraging victims of pornographic deepfakes to seek help.

The psychological toll on victims is immense. One individual became unable to attend school after fake images of them were spread online along with their school's name, paralyzed by the fear that "a stranger might be waiting in front of the school." Kazuna Kanajiri, chairperson of PAPS, emphasized, "The fear experienced by victims is immeasurable," and strongly urged those affected to reach out for support.

How to Seek Help

PAPS provides crucial support for victims. They can be contacted by:

• Phone: 050-3177-5432 (in Japanese)
• Email: paps@paps-jp.org

(This report is based on an original article by Yuki Machino, Business News Department, The Mainichi Shimbun.)