On July 10, 2023, the Cyberspace Administration of China (CAC), joined by six other central government regulators, issued finalized measures aimed at regulating the provision of generative AI services, such as ChatGPT, to the public of mainland China. (Interim Measures for the Management of Generative Artificial Intelligence (AI) Services.) Previously, in April, the CAC released a draft of these measures to solicit comments from the public.
The finalized measures, which still contain the word “interim” in their title, were jointly issued by the following regulators in addition to the CAC:
- National Development and Reform Commission.
- Ministry of Education.
- Ministry of Science and Technology.
- Ministry of Industry and Information Technology.
- Ministry of Public Security.
- National Radio and Television Administration.
According to the measures, China’s regulatory objective concerning generative AI includes promoting its healthy development and regulated application, as well as “safeguarding national security and social public interests” and “protecting the lawful rights and interests of citizens, legal persons, and other organizations.” (Finalized Interim Measures art. 1.)
The measures apply to the use of generative AI technology to provide services that generate any text, image, audio, video, or other such content to the public within mainland China. The development and application of generative AI technology that does not provide services to the public are explicitly excluded from the measures’ scope of application. (Art. 2.)
“Generative AI technology” is defined by the measures as “models and related technology that have the ability to generate text, image, audio, videos, or other such content.” A “generative AI service provider” is “any organization or individual that uses generative AI technology to provide generative AI services (including providing such services through the provision of a programmable interface or other means).” (Art. 22.)
Article 6 of the draft measures required providers to apply to the CAC for a security assessment before providing generative AI services to the public. This security assessment is addressed in article 17 of the finalized measures, which requires those providing generative AI services “with public opinion attributes or social mobilization capabilities” to apply to the CAC for a security assessment in accordance with relevant state provisions. The requirements of algorithm filing under the Algorithmic Recommendation Provisions (unofficial English translation) would also apply. (Art. 17.)
The language in the draft measures released in April holding providers of generative AI services responsible for the legitimacy of the source of any pretraining data or optimization training data used for their generative AI product has been removed. Instead, the measures require service providers to “employ effective measures to improve the quality of training data and to enhance the data’s veracity, accuracy, objectivity, and diversity.” (Art. 7.)
The language in the draft measures specifically requiring generative AI service providers to ask their users to register their real identities has also been deleted from the finalized measures. The new article 9 now states:
Regarding the obligations of protecting users’ input information and usage records, the measures proscribe not only unlawfully retaining input information that can identify a user or providing users’ input information to others, but also “collecting unnecessary personal information.” (Art. 11.)
The finalized measures do not include the fine of up to 100,000 yuan (about US$13,920) that the CAC and other relevant competent authorities could have imposed for violations under article 20 of the draft measures. A new article has been added in the finalized measures that addresses providing generative AI services that originate from outside the territory of mainland China to the public in China. If the provision of such services is against Chinese laws, regulations, or the generative AI measures, “the national cyberspace authority shall notify the relevant bodies to employ technical measures and other necessary measures to deal with the matter.” (Art. 20.)
The finalized generative AI measures will take effect on August 15, 2023. (Art. 24.)
Laney Zhang, Law Library of Congress
July 19, 2023
Read more Global Legal Monitor articles.