China Censors AI-Altered Videos Amid Concerns Over Cultural Integrity
On Monday, China’s National Broadcasting and Television Administration (NRTA) announced the removal of around 8,000 AI-altered videos from the state-run China Central Television (CCTV) platform. The videos had been flagged for “distorting, parodying, or obscenely depicting” classic Chinese films, television dramas, and animated works.
In a somewhat ironic twist, the state-owned Global Times reported on how two branches of the Communist government commended each other for their efforts in censorship. The NRTA has instructed major online audiovisual platforms to enhance their oversight responsibilities, ramp up daily monitoring, and prioritize the elimination of non-compliant AI-altered videos that alter or distort classical works based on renowned Chinese novels, historical events, and heroic figures.
Additionally, platforms are required to erase various forms of inappropriate animated content to foster a healthier online audiovisual environment, according to CCTV. NRTA’s initiative specifically targets three categories of content: AI-altered videos that greatly distort the essence or character portrayal of original works, content featuring graphic violence or vulgarity, and those that misinterpret Chinese cultural elements, leading to a skewed understanding of history.
The Global Times highlighted an example involving Lin Dae-woo from the 18th-century novel Dream of the Red Chamber, noting that the character had been inappropriately depicted as a “violent combatant.” Interestingly, Lin is often regarded as a sensitive figure, resembling a distinctly Chinese version of Hamlet. In the altered video, her character overcomes challenges through physical confrontations, showcasing the CCP’s apparent lack of humor regarding classical literature.
Furthermore, a change to the storyline of Empresses in the Palace, where characters are dropped into modern gunfight scenarios, was described as completely altering public perceptions of those figures. The series itself has been quite popular, depicting life in the imperial harem during the early 18th century. However, it seems that viewers found humor in envisioning the characters wielding modern firearms, something that evidently doesn’t sit well with the current government’s standards.
Starting from late 2024, as the censorship campaign against AI videos commenced, the NRTA expressed concerns that these humorous edits could undermine the public’s understanding of traditional culture and violate intellectual property rights. They warned platforms that they must intensify their reviews of AI-generated content and conduct self-assessments to avoid repercussions. Stricter regulations were enforced in March 2025.
In response, social media platforms have begun implementing measures to label or remove AI-generated text and videos. Regulators have voiced worry about AI’s potential to create convincing human expression, hinting at a belief that even with clear labeling, reimagining cherished stories could muddle public perception.
