On 1 November 2023, the Bletchley Declaration, the world’s first international agreement on artificial intelligence (“AI”), was issued at the AI Safety Summit held in the United Kingdom. The declaration reflects the global consensus on the promotion and regulation of AI. China is one of the 29 signatories of the declaration. As a major country in filing AI patents in the world, China not only promulgated “Interim Measures for the Administration of Generative Artificial Intelligence Services” but also announced a “Global Artificial Intelligence Governance Initiative” in the second half of 2023.
Interim Measures for the Administration of Generative Artificial Intelligence Services (“Interim Measures”)
On 10 July 2023, the Cyberspace Administration of China, National Development & Reform Commission, Ministry of Education, Ministry of Science & Technology, Ministry of Industry & Information Technology, Ministry of Public Security, National Radio and Television Administration jointly promulgated the Interim Measures which has become effective since 15 August 2023. The Interim Measures contain 24 articles, mainly including general principles, technology development and governance, service specifications, supervision and legal responsibilities.
Generative AI technology refers to models and related technologies that can generate text, pictures, audio, video and other content. Under the Cybersecurity Law of the People's Republic of China, Data Security Law of the People's Republic of China, Personal Information Protection Law of the People's Republic of China and Science and Technology Progress Law of the People’s Republic of China, the Interim Measures stipulate that generative AI services should be subject to inclusive, prudent and classified supervision. It also imposes the following requirements on providing or using the generative AI service:1. Do not commit acts that endanger national security and interests, harm the country's image, promote terrorism, ethnic hatred, ethnic discrimination, spread violence, obscenity, and false and harmful information, or other activities prohibited by laws and regulations;2. In the process of algorithm design, training data selection, model generation and optimization, and service provision, effective measures must be taken to prevent discrimination based on ethnicity, belief, nationality, region, gender, age, occupation, health, etc.;3. Respect intellectual property rights and business ethics, keep business secrets, and shall not take advantage of algorithms, data, platforms, etc. to implement monopoly and unfair competition;4. Respect the legitimate rights and interests of others, and shall not endanger the physical and mental health of others, or infringe on others’ rights to portrait, reputation, honor, privacy and personal information;5. Based on the characteristics of the service type, take effective measures to improve the transparency of generative AI services and improve the accuracy and reliability of generated content.
In terms of technology development and governance, the Interim Measures encourage the innovative application of generative AI technology in various industries and encourage the innovation of generative AI basic technologies, such as algorithms, frameworks, chips and supporting software platforms, international exchanges and cooperation, and the formulation of relevant international rules. In addition, the Interim Measures impose the following requirements on generative AI service providers to ensure that their AI materials are legal: 1. During the training and optimization process, it is necessary to ensure that the data and basic models have legal sources, do not infringe the intellectual property rights of others, and that the personal information involved and sampling is legal.2. Clear labelling rules should be formulated for data labelling in the generative AI technology research and development process, and labelling personnel should be trained to ensure labelling quality.
In terms of service specifications, the Interim Measures clarify that generative AI service providers should assume the responsibility of producers of online information content and have the following obligations towards users who register for their generative AI services:1. Clarify and disclose the applicable groups, occasions, and purposes of its generative AI services, and prevent minors from becoming addicted;2. Fulfill the obligation to protect personal information and correct and delete personal information according to user needs;3. Label generated content such as pictures and videos following “Provisions on the Administration of Deep Synthesis of Internet-based Information Services”; 4. Provide safe, stable and continuous generative AI services to ensure normal use of users.
In terms of supervision and legal responsibilities, the Interim Measures stipulate that China will formulate corresponding classified and hierarchical supervision rules or guidelines on the characteristics of generative AI technology and its service applications in relevant industries and fields. Generative AI services with public opinion attributes or social mobilisation capabilities should undergo security assessments and perform algorithm filing, change, and cancellation filing procedures per Provisions on the Administration of Algorithm-generated Recommendations for Internet Information Services. Users of generative AI services have the right to complain and report illegal generative AI services. Anyone who violates the Interim Measures shall be punished by laws such as the Cybersecurity Law of the People’s Republic of China, Data Security Law of the People’s Republic of China, Personal Information Protection Law of the People’s Republic of China, and Science and Technology Progress Law of the People’s Republic of China. If there are no legal provisions, punitive measures include warnings, notifications of criticism, orders to make corrections within a time limit, or orders to suspend the provision of relevant services. For generative AI services originating from overseas that violate Chinese laws or regulations, the Interim Measures stipulate that the national cybersecurity and informatization department shall notify relevant agencies to take technical measures and other necessary measures to deal with them.
The Interim Measures are regulatory provisions for generative AI services. It reflects China’s attitude toward promoting balanced governance in AI governance. The Artificial Intelligence Act (“EU Act”), passed by the European Parliament on 14 June 2023, is a bill that focuses on regulation while taking into account promotion. Compared with the EU Act, the Interim Measures seems to be slightly biased towards promotion, because its provisions are mainly based on normative guidance, lack of targeted specific regulations, weak enforceability and insufficient penalties. For example, according to Article 21 of the Interim Measures, in addition to being based on existing data security and network security laws, penalties for violations do not involve any more severe administrative penalties.
Global Artificial Intelligence Governance Initiative (“Initiative”)
On 18 October 2023, at the opening ceremony of the 3rd “Belt and Road” International Cooperation Summit Forum, China’s President Xi Jinping announced the “Global Artificial Intelligence Governance Initiative”. The initiative proposes to the world China’s following initiatives to jointly promote AI governance:1. The AI development should adhere to the people-centered concept and be based on ensuring social security and respecting human rights and interests. Its research and application in tackling global challenges such as climate change and biodiversity conservation should be supported.2. AI products and services provided to other countries should respect the sovereignty and laws of other countries and be governed by the laws of other countries. AI technology must not be used to manipulate public opinion, spread false information, interfere in other countries’ internal affairs, and endanger other countries’ sovereignty.3. The AI development should abide by applicable international law and be consistent with the common values of peace, development, fairness, justice, democracy, and freedom for all mankind, and jointly prevent and combat the misuse and abuse of AI by terrorism, extremist forces, and transnational organized criminal groups. In the military field, all countries, especially major countries, should develop and use AI technology prudently and responsibly.4. Countries’ AI development should adhere to the principles of mutual respect, equality and mutual benefit. The world should share AI knowledge and make AI technologies available to the public under open-source terms, but should not use technology monopolies and unilateral coercive measures to create barriers to development and maliciously block the global AI supply chain.5. Create auditable, supervised, traceable, and trustworthy AI technology by establishing a risk-level testing and evaluation system, and continuously improving the interpretability and predictability of AI, improving the authenticity and accuracy of data, and ensuring that artificial intelligence is always under human control.6. Formulate laws and regulations to protect personal privacy and data security in AI research and development and application, and prevent theft, tampering, leaking and other illegal collection and use of personal information.7. Avoid prejudice and discrimination against different or specific ethnic groups, beliefs, countries, genders, etc. in the process of data acquisition, algorithm design, technology development, product development and application.8. Adhere to ethics first, and establish a science and technology ethics review and supervision system.9. To achieve broad consensus in the field of international AI governance through dialogue and cooperation.10. Develop the development and application of relevant technologies for AI governance, and support the use of AI technology to prevent AI risks.11. Support discussions on the establishment of an international AI governance institution within the framework of the United Nations to coordinate major issues in international AI development, security and governance.
According to the initiative, China emphasises national interests and collective will for AI governance. The United States and China are both major countries in AI technology. Recognising that technology and governance are two sides of the same coin for AI, the United States has recently taken a series of AI governance measures, including signing a presidential executive order on 30 October 2023, requiring that developers of AI systems that pose a risk to the United States’ national security, economy, public health or security must share safety test results with the United States’ government under the Defense Production Act before releasing AI products to the public, and announced the creation of an AI regulator under the National Standards and Technology Research Institute affiliated with the U.S. Department of Commerce on 1 November 2023. Compared with China, what the United States values in AI governance is the government’s control over the free economy.
The leading position of China and the United States in AI technology has led to their dominant position in global AI governance. On 16 November 2023, when the Chinese leader Xi Jinping and the United States’ President Joe Biden met, they agreed to establish an intergovernmental dialogue on AI. It is hoped that the two countries can work hard to promote the establishment of efficient global AI regulatory agencies as soon as possible as the speed and method of development of AI technology can exceed human imagination.