News On Japan

Clothes Removed by Deepfake AI in 30 Seconds

OSAKA, Jul 16 (News On Japan) - Sexual deepfakes, created using generative AI, are rapidly emerging as a new form of digital abuse, with cases increasing across Japan. Without their knowledge, individuals—especially minors—are finding their photos misused to produce sexually explicit images or videos, often in under a minute.

One disturbing trend involves photos of girls in school gym uniforms being digitally altered to depict nudity. In many instances, graduation album photos are exploited. "It looks like the clothes are stripped off," said one source familiar with how easily these AI tools operate.

Unlike earlier years when celebrities were mainly targeted, more recent victims include ordinary middle and high school students—and even elementary school children. According to a study by Hiiragi Net, which monitors digital abuse, there were 252 confirmed cases of sexual deepfakes involving minors between March and June this year. Many of the images were full-body nudes or manipulated to appear as though the children were engaging in sexual acts.

Graduation season in March and April appears to be a peak period for these abuses. The moment schools distribute albums, perpetrators begin uploading the photos to apps or websites where generative AI strips clothing and alters body features. Some photos are even shared on social media platforms with tags like "graduation album," and sections labeled “sotsu” (short for sotsugyō, or graduation) containing multiple AI-altered images.

In some cases, classmates themselves have requested nude edits of fellow students by submitting original photos to these platforms. The creation process is alarmingly easy—users simply upload a headshot, choose options like body size, click generate, and within about 30 seconds, a realistic fake nude image is produced.

One journalist who tested a service described being shocked at the realism: "If I hadn’t been told, I wouldn’t have realized the photo had been created by AI. It looked completely natural."

Hiiragi Net’s representative Nagamori warned that once images are created and shared online, individuals have little control. "Graduation photos or event pictures are often beyond personal management. In truth, full self-protection is no longer realistic. The government must implement proper countermeasures."

In response, some photo album publishers are beginning to change their practices. One Osaka-based company said it now considers omitting identifying details, or separating names from faces. However, security remains a major concern. In April, a printing company in Sendai reported a cyberattack that may have leaked the personal data of 173,000 students nationwide. The company says it now uses top-tier security systems to prevent recurrence.

"We take pride in creating graduation albums as a cherished part of student life. But once the product leaves our hands, we have no control," said a representative of the company.

Beyond graduation albums, other photos have also been targeted. In some cases, ID and password leaks from nursery school photo sites led to images being taken and manipulated. Even wedding photos shared on social media have reportedly been turned into deepfakes.

Concerns about regulation are mounting. While Japan currently lacks laws specifically targeting sexual deepfakes, some local governments are taking action. Tottori Prefecture introduced an ordinance in April banning the creation and distribution of child pornography using generative AI, with penalties of up to 50,000 yen and mandatory deletion orders. From next month, those who fail to comply could face further fines or public name disclosure.

Experts note, however, that national legislation is still absent, and global enforcement is complicated—particularly when the offenders or hosting services are overseas. In contrast, the European Union is moving to regulate AI more broadly, including imposing penalties on companies that produce tools used in human rights violations.

"In Japan, regulation must be implemented urgently before the harm spreads further," said Nagamori. "This is a terrifying new era where any child—or adult—could have their image stolen, stripped, and turned into a deepfake in seconds."

Source: KTV NEWS

News On Japan
POPULAR NEWS

An Idemitsu Kosan crude oil tanker has safely passed through the Strait of Hormuz, becoming the first vessel bound for Japan to do so since attacks on Iran heightened tensions in the region and effectively disrupted maritime traffic.

Japan’s Golden Week holiday period got fully underway on April 29, drawing large crowds to major tourist destinations and airports, where long lines formed as overseas travel surged.

A series of sightings involving unusually large brown bears in Hokkaido has heightened concerns among local residents, with one 330-kilogram animal captured in Tomamae and another 280-kilogram bear attacking a hunter in Shimamaki.

Full-scale Golden Week travel began on April 29, with Chubu Centrair International Airport experiencing its busiest outbound travel day of the holiday period. The airport was crowded from the morning with vacationers heading overseas.

Electricity and gas bills for usage in May will rise slightly in Japan, with the impact of tensions involving Iran expected to appear in utility charges from June onward. Larger increases could follow in subsequent months.

MEDIA CHANNELS
         

MORE Web3 NEWS

Generative AI chatbot ChatGPT would have ranked first in this year's entrance examinations for both the University of Tokyo and Kyoto University, according to research by AI startup Life Prompt.

画像から動画へという技術は、静止コンテンツをより魅力的な動的コンテンツへと変換する手段として急速に普及している。

Japanese financial services group Orix Corp has announced a slate of senior leadership changes.

The Japanese government held an emergency meeting with the Bank of Japan and executives from the country's three megabanks to discuss potential risks posed by the new artificial intelligence model 'Claude Mythos' to Japan's financial system.

Japan is not like other markets. Brands that try to copy-paste their Western influencer strategies here will struggle. Japanese consumers have unique platform preferences, cultural expectations, and engagement patterns that require a localized approach.

Businesses worldwide are grappling with a growing concern: will artificial intelligence take away human jobs? Yet in the United States, recent developments suggest a more complex reality. Job openings for software developers have been rising since around November 2025, according to U.S. employment site Indeed.

LINE Yahoo announced a new 'AI Agent' service on April 20, aiming to provide an AI assistant that users can rely on in their daily lives. The service will be introduced through the LINE app and Yahoo! JAPAN platform.

VICTORIA, Seychelles, March 31, 2026 - Global crypto trading platform BYDFi will mark its 6th anniversary with a month-long celebration beginning on April 1, 2026, highlighting BYDFi’s evolution into an all-in-one crypto trading platform built on a CEX + DEX dual-engine model.