China has issued draft security regulations for companies providing generative artificial intelligence (AI) services, which include restrictions on data sources used to train AI models.
On Wednesday, October 11, the proposed regulations Released By the National Information Security Standardization Committee, which includes representatives of the Cyberspace Administration of China (CAC), the Ministry of Industry and Information Technology and law enforcement agencies.
Generative AI, as demonstrated by OpenAI’s ChatGPT achievements, gains the ability to perform tasks by analyzing historical data and creating new content, such as text and images, based on this training.
The committee recommends a security assessment on content used to train generative AI models that is publicly accessible. Content exceeding “5% in the form of illegal and harmful information” will be blacklisted. This category includes content that calls for terrorism, violence, subversion of the socialist system, damage to the country’s reputation, and actions that undermine national cohesion and societal stability.
The draft regulations also emphasize that censored data on the Chinese Internet should not serve as training material for these models. The development comes just over a month after regulatory authorities granted permission to several Chinese technology companies, including prominent search engine Baidu, to offer generative AI-based chatbots to the general public.
Since April, the CAC has consistently communicated its requirements for companies to submit security assessments to regulators before offering AI-powered generative services to the public. In July, the Cyberspace Regulatory Authority issued a set of guidelines governing these services, which industry analysts noted were significantly less burdensome than measures proposed in an initial draft in April.
Related: Biden is considering tightening controls on artificial intelligence chips on China through third parties
The recently revealed draft security terms require that organizations involved in training these AI models obtain explicit consent from individuals whose personal data, including biometric information, is used in the training. In addition, the Guidelines include comprehensive instructions on preventing intellectual property violations.
Countries around the world are grappling to create regulatory frameworks for this technology. China considers artificial intelligence a field in which it aspires to compete with the United States, and has set its ambitions to become a global leader in this field by 2030.
magazine: “AI has killed the industry”: EasyTranslate CEO on adapting to change