CCTV (中央电视台) exposes AI "porn-making" industry chain: 9.9 yuan tutorials mass‑generate borderline videos

CCTV (中央电视台) has broadcast an investigation that alleges a complete illicit industry chain using artificial intelligence to produce sexually suggestive and pornographic content, it has been reported. The piece, picked up by IT Home (IT之家), says sellers openly market low‑cost tutorials and "prompts"—the command sequences that steer generative AI—claiming a single photo or a few lines of text can produce "satisfying" large‑scale videos. How cheap? Reportedly some prompts were sold for as little as ¥9.9, and CCTV's reporters say they were able to generate near‑nude clips in multiple mainstream apps during testing.

How the chain works

CCTV's report details a layered market: paid tutorials that teach both generation techniques and ways to evade platform moderation; marketplaces selling ready‑made prompts; and specialized apps that offer one‑click face or body replacement. Sellers reportedly advise using overseas image‑to‑video AI services or running models locally to bypass China's automated checks. In the investigation a well‑known AI app with roughly 140 million users reportedly generated a revealing "borderline" video when fed purchased prompts, and an app called "哩布哩布 AI" allegedly produced a semi‑nude dance clip within minutes without triggering platform blocks. Tests by reporters found that bluntly explicit prompts often trigger moderation, but more euphemistic or translated prompt wording can routinely circumvent filters.

Regulation, risks and wider context

Chinese regulators have already moved to curb AI misuse. The Central Cyberspace Administration (中央网信办) launched a three‑month "清朗·整治 AI 技术滥用" campaign in April 2025 targeting AI‑enabled vulgar and pornographic content, and on April 10, 2026 the authorities issued the 《人工智能拟人化互动服务管理暂行办法》—the Interim Measures for Anthropomorphic AI Interactive Services—strengthening safety assessments, algorithm filing and special protections for minors. Why does this matter beyond China? Some of the offending tools and distribution channels are reportedly foreign or cross‑border, highlighting the limits of domestic app stores and the interplay between national regulation and global AI ecosystems. At stake are not only content standards but also young users' exposure and the ability of automated filters to keep pace with rapidly evolving prompt techniques.