China Greenlights 487 Deepfake Algorithms from Baidu, Alibaba, and Tencent
China has approved 487 deepfake algorithms from digital titans such as Baidu, Alibaba, and Tencent. Regulations demand that AI technologies be registered with the government. Foreign firms, such as HP and Yum China, have also had their algorithms approved.
These algorithms are developed by prominent Chinese technology businesses such as Baidu, Alibaba Group Holding, and Tencent Holdings, as well as foreign companies such as Hewlett-Packard (HP).
The Cyberspace Administration of China (CAC) has published the second-largest list of approved algorithms since the regulations were established in January 2023. Among the endorsed algorithms are those built into products like Baidu's portrait image diffusion generator in its cloud gallery app Yike and a search algorithm in Tencent's WeChat.
Alibaba, the parent company of the South China Morning Post, has registered an algorithm for document creation in its enterprise tool DingTalk. Additionally, its research arm Damo Academy has developed a multimodal algorithm tailored for corporate clients to produce images and videos.
These actions are consistent with the Administrative Provisions on Deep Synthesis for Internet Information Services, which regulate deepfake technologies that use deep learning and augmented reality to generate text, photos, audio, video, or virtual scenes that resemble real-life things. Failure to register an algorithm with the CAC may result in removal from domestic app stores.
The CAC has already issued six other whitelists, the most recent batch of 492 in June being the most comprehensive. The lists have been gradually expanding, with the first list in June 2023 including only 41 algorithms.
Cai Peng, a partner at Zhong Lun Law Firm in Beijing, believes that the growing size of the lists reflects a more efficient process between the regulator and applicants. The application process normally consists of submitting and revising documentation in accordance with the CAC's standards.
Notable contributions in the recent batch include a healthcare knowledge algorithm for Douyin (the Chinese counterpart of TikTok), a song generator from Microsoft's AI spin-off Xiaoice, and a character dialogue generator for NetEase's digital translation of the famous party game Werewolf.
Foreign brands have also gained slots on the list, with HP's computer assistant algorithm and a text generator developed by Yum China, a subsidiary of American fast-food conglomerate Yum! Brands, making the cut for usage in KFC, Pizza Hut, and other restaurants.
China has emerged as a frontrunner in AI regulation, mandating developers to register relevant technologies with the government. GenAI models must be registered with the CAC before public release, as per a regulation implemented in August 2023. The regulator has already approved a combined total of 188 GenAI models across two lists.
China approves 487 deepfake algorithms from tech giants like Baidu, Alibaba, and Tencent
Regulations require registration of AI technologies with the government
Foreign companies, including HP and Yum China, also have algorithms approved
Source: SCMP