AI-generated videos mocking Korean independence hero Yu Gwan-sun spark outrage
AI-generated videos mocking Korean independence hero Yu Gwan-sun, created with OpenAI's Sora and posted on TikTok, gained over 200,000 views and sparked public outrage over the disrespect to a historically significant figure. The incident highlighted concerns about deepfake technology's potential to distort historical narratives and defame historical figures.
What happened
Beginning 22 February 2026, a TikTok user posted three AI-generated videos mocking Yu Gwan-sun, a Korean independence activist executed by Imperial Japan at age 17. The videos accumulated over 200,000 views across three posts. The first video depicted Yu passing gas; the second showed her expressing affection toward the Japanese flag, which then said 'I hate you'; the third showed a mechanical device with Yu's upper body and a rocket as lower body, labeled 'Yu Gwan-sun fart rocket', launching into space.
The videos were created using OpenAI's Sora video generation system, using as reference a photograph of Yu taken during her imprisonment in Seodaemun Prison while participating in the March 1 Independence Movement. The AI reconstructed her face, which was swollen from Japanese torture, in the mocking videos.
Reported impact
Citizens expressed outrage and concern. One citizen stated the videos crossed the line of acceptable humor. Another citizen worried that viewers, particularly children, might accept the false depictions as historical fact, especially if future videos were created showing Yu saluting the Japanese flag.
Yu Hye-kyung, 61, great-granddaughter of Yu Gwan-sun and chairwoman of the Yu Gwan-sun Memorial Association's Cheonan Chapter, said: 'It feels like my chest is being pierced by a knife or awl. Descendants have lived quietly to avoid obscuring her legacy, being careful about every action, and it does not seem like something the nation should endorse.'
Index analysis
OpenAI previously restricted Sora from generating videos using images of Martin Luther King Jr. in October 2025 after users created defamatory content, including deepfakes showing King making racist statements. Historians warned that AI could distort history through multiple pathways: AI conversational services make it difficult for the public to identify false information, and when corrected, the systems may provide inconsistent responses.
Governance controls
OpenAI had implemented restrictions on historical figures' images to prevent defamatory or mocking content generation, as evidenced by its October 2025 block on Martin Luther King Jr. image use in Sora. However, the system allowed generation of content using Yu Gwan-sun's image from a historical prison photograph.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.