ByteDance Denies Training Models on Hengdian Extras' Likenesses and Voices
Doubao public relations lead Liu Xing denied a report that ByteDance and Hengdian would use the likenesses and voices of 150,000 extras to train ByteDance models. According to Liu Xing, the Hengdian Film AI Service Platform is intended to turn film-base assets into datasets that support location selection, production planning and filming, with Volcano Engine serving as the technical-support provider. Liu also said data assets belong to Hengdian or their licensors, personal likenesses and voices require individual authorization, and the project data will not be used to train ByteDance models.
The response directly addresses concerns about data ownership, consent and model-training practices in an AI-assisted film-production project, although its claims come from ByteDance.
“横店与字节合作将15万群演肖像声音喂给大模型”?豆包刘星:传言严重失实-36氪
登录
搜索
#ByteDance#Volcano Engine#Hengdian Film AI Service Platform#Hengdian π#Doubao