Four in Japan Arrested for Selling AI-Generated Pornographic Posters, Earning Millions

Japanese police arrested four suspects, aged between their 20s and 50s, for using free generative AI software to create nude images of non-existent women, printing them as posters and selling them on online auction sites. This marks Japan's first arrest for AI-generated obscene materials, with one suspect earning approximately 10 million yen in a year.

Facts Restored

A report on April 15, 2025, showed that four individuals, including Tomohiro Mizutani, a 44-year-old retailer from Kitanagoya City, Aichi Prefecture, were arrested on suspicion of distributing obscene materials. Since October of the previous year, they had repeatedly sold AI-generated nude posters on online auction sites. Three admitted to the charges, while one partially denied them. Police confirmed this is the first time in Japan that AI-generated images have been classified as obscene materials, leading to arrests.

Mechanism Breakdown

The suspects used free generative AI software available online, inputting prompts such as "legs spread" to create nude images of adult women who do not exist in reality. When listing posters on the auction site, the lower half of the images was blurred to circumvent the platform's restrictions on obscene materials, but the actual shipped items were uncensored originals. Each poster sold for several thousand yen, and Tomohiro Mizutani earned approximately 10 million yen over about a year through this method.

This operation exploited the low-barrier nature of generative AI: no professional equipment or real models were needed, and content could be mass-produced using only publicly available software. The business logic relied on bypassing platform reviews while meeting specific buyer demands, with blurring serving as a direct means to evade rules.

Industry Impact

For developers, the misuse of free AI tools for profit may accelerate platforms' efforts to enhance content filtering and usage restrictions, increasing compliance costs. For enterprise users, similar incidents could reduce trust in generative AI services in commercial scenarios, affecting downstream applications such as advertising and design. In terms of competitive landscape, independent media coverage of this case has highlighted global concerns over malicious AI usage, potentially prompting more regions to assess whether existing laws cover AI-generated content.

Comparison and Precedents

A 2019 study by the Dutch company Sensity showed that approximately 96% of online deepfake videos were non-consensual pornographic content, with the majority targeting women. This case aligns with that data, revealing a common pattern where AI-generated pornographic content extends from real individuals to fictional images. Japan's action represents the first law enforcement precedent targeting AI-generated materials, distinct from previous discussions focusing only on deepfakes of real people.

Strategic Judgment

Based on the details of the current case, it is most likely that platforms will further tighten rules on listing AI-generated content, and more jurisdictions will test similar legal applications following Japan's approach. The incident reveals the actual execution path of generative AI in commercial chains: from prompt generation to blurring processing and then physical delivery, each step relies on existing technological and regulatory gaps. Law enforcement has, for the first time, included AI-generated outputs within the category of obscene materials, altering the previous regulatory framework that only targeted traditional production methods.

For developers, the misuse of free tools may prompt software providers to add watermarking or trackable marking features to reduce risks of being used for illegal sales. For enterprise users, the incident may raise the compliance review threshold for using generative AI, helping avoid involvement in similar legal disputes. In terms of competitive landscape, sustained independent media coverage may amplify public discussion on AI misuse, indirectly directing investment toward compliance-oriented projects.

Compared to historical precedents, this case differs from previous deepfake crimes targeting real photos, focusing instead on the commercialization of entirely fictional images, highlighting the need for laws to adapt to the characteristics of AI-generated content. The age range of the four suspects and the profit figures have been confirmed by multiple sources, and motives such as using the proceeds for living expenses have also been mentioned.