China has taken a significant step in regulating generative artificial intelligence (AI) services. The country just unveiled security regulations that outline restrictions on the data sources used for training these AI models.
The proposed regulations was released on Wednesday, Oct. 11, by the National Information Security Standardization Committee. Further, it brought together representatives from the Cyberspace Administration of China (CAC), the Ministry of Industry and Information Technology, and law enforcement agencies.
Security Evaluation for Publicly Accessible Generative AI Models
Generative AI, exemplified by OpenAI’s ChatGPT, gains task-performing abilities through historical data analysis, generating fresh content like text and images. The committee however, recommends a thorough security evaluation of content used to train publicly accessible generative AI models.
Also, content surpassing “5% in the form of unlawful and detrimental information” will be blocklisted. This category includes content advocating terrorism, violence, subversion of the socialist system, harm to the country’s reputation, and actions undermining national cohesion and societal stability.
The draft regulations underscore that data subject to censorship on the Chinese internet should not be utilized as training material for these models. This development further follows regulatory permission granted just over a month ago to various Chinese tech companies. These companies include Baidu, to introduce generative AI-driven chatbots to the public.
Consistent Regulatory Requirements
Moreover, since April, the CAC has consistently communicated the requirement for companies. This is to undergo security evaluations before introducing generative AI-powered services to the public. In July, the cyberspace regulator released guidelines governing these services, noted to be less burdensome compared to the measures proposed in the initial April draft.
Notably, the newly unveiled draft security stipulations mandate organizations training these AI models to obtain explicit consent from individuals. These include personal data, including biometric information, used for training. Additionally, the guidelines include comprehensive instructions on preventing infringements related to intellectual property.
Global Competition and China’s AI Ambitions
As nations worldwide grapple with establishing regulatory frameworks for AI, China sees AI as a domain in which it aims to compete with the United States. Notably, the country has set ambitious goals to become a global leader in AI by 2030.