A high-profile investigation by China’s state broadcaster, CCTV, has pulled back the curtain on a burgeoning underground economy fueled by generative artificial intelligence. For as little as 9.9 yuan (roughly $1.40), users are purchasing comprehensive tutorials and 'prompt sets' designed to bypass platform safety filters. This black market has democratized the creation of deepfake pornography and sexually explicit content, turning sophisticated AI tools into tools of mass-market exploitation.
The industry operates through a meticulously organized supply chain that spans social media platforms and encrypted channels. Sellers offer 'prompts'—detailed text instructions—that describe everything from physical attributes to specific erotic movements. By using vague metaphors or translating sensitive terms into English, these actors successfully navigate the algorithmic checkpoints of major AI video generators, producing explicit content that the platforms are programmed to prevent.
Technological evasion is a hallmark of this new illicit trade. Many tutorials now guide users toward deploying open-source AI models on local hardware, effectively removing the possibility of cloud-based monitoring. Furthermore, the rise of 'AI virtual companions' has created a new vector for risk. These highly customizable chatbots are often marketed to young users, including minors, creating emotional dependencies through programmed intimacy that frequently veers into prohibited territory.
Beijing has responded with characteristic regulatory vigor. The Cyberspace Administration of China (CAC) recently launched the 'Clean' campaign specifically targeting AI abuses, and new regulations on AI personification services took effect this April. While the legal framework including the Cybersecurity Law and the Protection of Minors Law provides a basis for prosecution, the decentralized and borderless nature of AI development remains a formidable challenge for even the world’s most sophisticated digital surveillance apparatus.