China's cyberspace regulator has recently escalated its campaign against AI-related misconduct, focusing on the second phase of the "Qinglang" (Clear and Bright) special operation. This phase targets pressing issues such as using AI to fabricate false information, spreading violent and vulgar content, impersonating others, and infringing on minors' rights. Local regulators have been directed to intensify enforcement, pushing online platforms to purge harmful AI-generated material and take legal action against offending accounts and applications, thereby curbing the misuse of AI-generated content and fostering a healthier, more orderly development of AI application services.

As of the latest count, the campaign has cleared over 5.61 million pieces of illicit information, penalized more than 49,000 accounts, and dealt with over 2,400 illegal websites and apps. During the operation, a slew of violating accounts and applications have faced consequences, and the regulator has now publicized several typical cases as a warning against future offenses.

AI "Tampering" with Classics and Generating "Digital Slop"

Accounts on major platforms, including a Weibo account tagged "天**小", a Douyin account "歪**国", a Kuaishou account "小*菌", and a WeChat account "咸*i", have been sanctioned for misusing AI to alter classic literary works like Romance of the Three Kingdoms and Journey to the West. These actions subverted the established character personas and fabricated low-quality, sensational plotlines to attract viewership, leading to their removal from the platforms per regulations.

AI-Generated False and Misleading Information

Accounts such as the Kuaishou account "云**视", Xiaohongshu account "有*头", Douyin account "7**员", and WeChat account "长**帆" were found to have used AI to create fake disaster scenes and rescue footage, fabricate trending social topics, promote superstitious content, and deploy AI digital avatars for deceptive marketing. These actions misled elderly users and stoked public panic, violating platform rules and resulting in their closures.

AI Impersonation and Counterfeiting of Others

The Douyin account "特**力", Kuaishou account "石*", Xiaohongshu account "潮**件", and WeChat account "M**勒" have been shut down for employing AI-driven "face-swapping" and "voice-cloning" technologies to mimic public figures, producing and distributing lowbrow dance clips and images of embraces or kisses. These tactics were used as a commercial ploy to generate revenue, which led to their legal and contractual penalties.

Spreading Violent and Vulgar Content

Accounts including the Bilibili account "带**着", Douyin account "波**C", Kuaishou account "虾**诡", and WeChat account "星**7" have been penalized for mass-producing AI-generated videos featuring horrific, eerie, and violent themes, prompting action from the authorities to clean up the digital environment.

AI Endangering Minors' Well-Being

The Xiaohongshu account "小**娃", WeChat account "豆*a", Douyin account "半**历", and Kuaishou account "若**雪" have also faced consequences for using AI to create and distribute content showing minors in pregnancy or fighting scenarios, which is inconsistent with their age. These accounts also churned out "toxic animations" and "cult-style" videos that mocked or distorted images of children, negatively impacting their psychological health, leading to their removal from the platforms.

AI "Managed" Accounts for Unscrupulous Online Activities

WeChat account "涛**客", Weibo account "A**师", and Douyin account "风*" have been dealt with for using AI "management" software to simulate real user operations and interactions, generating a flood of low-quality, repetitive content that fueled click-farming and other water-army tactics, violating platform guidelines.

Violations by AI Products and Applications

Several websites and platforms, including "N**t", "穿**元", "甜*圈", and "芋**o", have been flagged because their AI bot covers, descriptions, and generated dialogues included pornographic or other material harmful to minors' well-being, severely disrupting the online ecosystem. Those that refused to rectify the issues face legal measures, with steps such as warnings, fines, and removal or shutdown actions being taken as appropriate.

Looking ahead, the cyberspace authorities plan to build on the campaign's achievements by closely monitoring AI development trends and potential risks, implementing timely measures to curb the abuse and malicious use of AI technology, and continuing to foster a supportive environment for the innovative growth of artificial intelligence.