A former gaming programmer and five accomplices have been sentenced to prison in China for using generative AI to produce and sell obscene content, marking one of the most significant enforcement actions yet against AI-powered adult content creation in the country.
The Yuyao People’s Court in Zhejiang Province handed down sentences ranging from three years and one month to three years and ten months to the six defendants, who developed an AI platform specifically designed to bypass safety restrictions and generate explicit material for paying users. The ruling has already taken effect after none of the defendants appealed.
How the Operation Worked
The case centers on a defendant surnamed Zheng, a former programmer at a major Chinese gaming company who reportedly encountered an overseas AI sex-chat bot in July 2024. Inspired by what he saw, Zheng assembled a development team including a colleague and several contacts he met online.
By January 2025, the group launched their generative AI model on an overseas platform, charging users through multiple payment channels including cryptocurrency. The service allowed paying customers to select different characters for sexual conversations with AI bots and spend credits to generate explicit static or animated images.
The technical approach was straightforward but effective: the team used prompt engineering to “jailbreak” their base model, repeatedly rewriting system prompts, setting up character cards, and creating specific dialogue scenarios to circumvent the model’s built-in restrictions on adult content.
The scale of the operation was substantial. During its operational period, the platform generated more than 370,000 static images and over 6,000 animated videos. Users paid approximately $24,000 (about 170,000 yuan) for access before police opened an investigation in July 2025. All six defendants were arrested within two months.
The Legal Precedent
The case established important legal ground in how China treats AI-generated adult content. A central question was whether developers could be held criminally liable for “producing” obscene materials when users were the ones selecting characters and entering specific prompts.
The court ruled that despite user participation, the defendants exercised substantial control over the model’s output through the technical framework they designed. Their conduct had a direct causal link to the production of obscene content, making them criminally responsible.
A judge handling the case stated that “technological neutrality” cannot serve as a shield for illegal activity, and developers cannot avoid responsibility simply because AI generates the final content. The court also noted that because AI can produce content at massive scale with minimal effort, traditional quantity-based standards for sentencing may not be appropriate. Instead, factors like distribution scope, risk of wider dissemination, and actual harm should be considered.
China’s Broader Crackdown on AI Adult Content
This case arrives amid an escalating regulatory campaign against AI companions and adult content in China. The country maintains some of the world’s strictest laws regarding pornography and obscene material, with production, distribution, and sale of pornographic content carrying criminal penalties including imprisonment.
Under Chinese criminal law, producing, duplicating, publishing, selling, or disseminating obscene materials for profit can result in prison sentences up to life imprisonment depending on the severity. Even possession of obscene materials can result in administrative detention.
The AI companion space has come under particular scrutiny. Earlier this year, China’s first AI companion chat application case entered a second-stance trial after developers and operators were detained. That case sparked widespread debate about the boundaries of emotional AI services and where legitimate technology ends and illegal content begins.
The message from Chinese authorities is clear: jailbreaking AI models to generate adult content is not a technical loophole—it’s a criminal offense. For developers in the adult AI space, the risks of operating in or targeting Chinese users have never been higher.
Sources include https://www.globaltimes.cn/page/202609/1370112.shtml
