China has unveiled a set of draft security regulations that focus on companies providing generative artificial intelligence (AI) services, with a specific emphasis on the sources of data used for training AI models.
The National Information Security Standardization Committee released these proposed regulations on October 11. The committee comprises representatives from the Cyberspace Administration of China (CAC), the Ministry of Industry and Information Technology, and law enforcement agencies.
Generative AI, as demonstrated by technologies like OpenAI’s ChatGPT, gains the ability to perform tasks by analyzing historical data and generating new content, including text and images, based on this training.
The committee recommends conducting a security assessment on the data used to train publicly accessible generative AI models. Content exceeding “5% in the form of unlawful and harmful information” will be classified for blacklisting. This category includes content that promotes terrorism, violence, undermines the socialist system, harms the country’s reputation, and disrupts national unity and social stability.
The draft regulations also highlight that data subject to censorship on the Chinese internet should not be utilized as training material for these AI models. This announcement comes a little over a month after regulatory authorities granted permission to various Chinese tech companies, including prominent search engine Baidu, to introduce their generative AI-powered chatbots to the public.
Since April, the CAC has consistently required companies to provide security evaluations to regulatory bodies before launching generative AI-powered services for public use. In July, the cyberspace regulator issued a set of guidelines for these services, which industry experts noted were less stringent compared to the measures initially proposed in the April draft.
The newly introduced draft security provisions mandate that organizations involved in AI model training obtain explicit consent from individuals whose personal data, including biometric information, is used for training. Additionally, the guidelines offer comprehensive instructions on preventing intellectual property infringements.
Countries worldwide are grappling with the establishment of regulatory frameworks for generative AI technology. China views AI as a sector in which it aims to compete with the United States and has set ambitious goals to become a global leader in the field by 2030.
