The Implementing Opinions encompass 23 initiatives across nine areas, aiming to achieve several objectives by 2023, including: to grow the data factor market to CNY200 billion in size, to successfully pilot the national basic systems for data, and to nurture clusters within the data service industry. Moreover, the Implementing Opinions introduce pivotal initiatives as follows: leading the way in implementing data property rights and revenue-distribution systems, accelerating the realization of the data asset value, fostering the growth of the data factor market, vigorously developing the data service industry, initiating the program to pilot the implementation of the basic systems for data, emphasizing efforts to enhance data classification and grading protection, implementing guidelines for data classification and grading in industries like autopilot, healthcare, industry, finance and transportation, strengthening research, development and application of security and guarantee technologies in respect of data security monitoring, encrypted transmission, access control, data desensitization, privacy computing, etc.

In recent years, generative AI technology has seen rapid development, ushering in new opportunities for economic and social advancement. Yet, it has also given rise to concerns, such as the spread of false information, infringement of personal data rights, data security issues, and bias and discrimination. The imperative to balance the development and security of generative AI has garnered significant attention. The introduction of the Measures is not only an essential requirement to promote the healthy development of generative AI, but also a practical necessity to mitigate risks associated with generative AI services.

The Measures state that China, guided by the principles of placing equal emphasis on development and security and combining innovation promotion with law-based governance, will implement effective measures to encourage the innovation and development of generative AI, and take an approach of inclusion, prudence, and classification and grading in regulating generative AI services. The Measures outline the general requirements for providing and using such services, and present specific measures to drive the development of generative AI technology along with detailed directives for activities such as training data processing and data labeling. They also establish standards for generative AI services, specifying that service providers must employ effective measures to prevent users who are minors from overreliance on or addiction to generative AI services, and should label images, videos, and other generated content in accordance with the Administrative Provisions on Deep Synthesis in Internet-based Information Services(the "Administrative Provisions on Deep Synthesis"), and promptly take action against any identified illegal content. Furthermore, the Measures institute systems for security assessment, algorithm filing, and handling of complaints and reports, while also delineating the legal liability.

According to the CAC officials, fostering the growth and governance of generative AI services necessitates collaboration among the government, businesses, civil society, internet users, and other pertinent parties, and this collective effort aims to promote the healthy development of generative AI, ensuring generative AI technology delivers greater benefits to the public.

Compared to the Draft, the Measures further clarify and narrow down the scope of   application: providing generative AI services to the public within the   People's Republic of China. When it comes to determine whether the Measures   apply, it's essential to consider three criteria: "within the territory   of China," "to the public," and "providing   services."

The   Measures establish detailed service standards for generative AI service   providers, the applicable entities are generative AI service providers, which   refer to organizations and individuals that utilize generative AI technology to   provide generative AI services (including providing such services through the   provision of a programmable interface or other means). It should be   emphasized that, unlike the Administrative Provisions on Deep Synthesis, the   Measures do not distinguish between technology supporters and service   providers, and they explicitly state that those "offering services through   the provision of a programmable interface or other means" are subject to   the Interim Measures. For instance, technology supporters who have completed   filings for their algorithms, published on June 24, 2023, are all providing services   by means of API, so it can be seen that both technical supporters and service   providers are subject to the same regulatory requirements under the Measures. 

The   Measures also explicitly express support for and encouragement of the   development of generative AI services, at the application layer, emphasis is   put on promoting the creation of positive and wholesome high-quality content,   while, at the technical and fundamental layers, the focus primarily lies in   accelerating the development of infrastructure (arithmetic, data, etc.) and   fostering equal international cooperation and participation in the   development of international standards.