Comprehensive governance measures
2025-02-06
While adopting technological controls, we should formulate and refine comprehensive AI safety and security risk governance mechanisms and regulations that engage multi-stakeholder participation, including technology R&D institutions, service providers, users, government authorities, industry associations, and social organizations.
5.1 To implement a tiered and category-based management for AI application.
We should classify and grade AI systems based on their features, functions, and application scenarios, and set up a testing and assessment system based on AI risk levels. We should bolster enduse management of AI, and impose requirements on the adoption of AI technologies by specific users and in specific scenarios, thereby preventing AI system abuse. We should register AI systems whose computing and reasoning capacities have reached a certain threshold or those are applied in specific industries and sectors, and demand that such systems possess the safety protection capacity throughout the life cycle including design, R&D, testing, deployment, utilization, and maintenance.
5.2 To develop a traceability management system for AI services.
We should use digital certificates to label the AI systems serving the public. We should formulate and introduce standards and regulations on AI outputlabeling, and clarify requirements for explicit and implicit labels throughout key stages including creation sources, transmission paths, and distribution channels, with a view to enable users to identify and judge information sources and credibility.
5.3 To improve AI data security and personal information protection
regulations.
We should explicate the requirements for data security and personal information protection in various stages such as AI training, labeling, utilization, and output based on the features of AI technologies and applications.
5.4 To create a responsible AI R&D and application system.
We should propose pragmatic instructions and best practices to uphold the people-centered approach and adhere to the principle of developing AI for good in AI R&D and application, and continuously align AI’s design, R&D, and application processes with such values and ethics. We should explore the copyright protection, development and utilization systems that adapt to the AI era and continuously advance the construction of highquality foundational corpora and datasets to provide premium resources for the safe development of AI. We should establish AI-related ethical review standards, norms, and guidelines to improve the ethical review system.
5.5 To strengthen AI supply chain security.
We should promote knowledge sharing in AI, make AI technologies available to the public under open-source terms, and jointly develop AI chips, frameworks, and software.We should guide the industry to build an open ecosystem, enhance thediversity of supply chain sources, and ensure the security and stability of the AI supply chain.
5.6 To advance research on AI explainability.
We should organize and conduct research on the transparency, trustworthiness, and error-correction mechanism in AI decision-making from the perspectives of machine learning theory, training methods and human-computer interaction. Continuous efforts should be made to enhance the explainability and predictability of AI to prevent malicious consequences resulting from unintended decisions made by AI systems.
5.7 To share information, and emergency response of AI safety
risks and threats.
We should continuously track and analyze security vulnerabilities, defects, risks, threats, and safety incidents related to AI technologies, software and hardware products, services, and other aspects. We should coordinate with relevant developers and service providers to establish a reporting and sharing information mechanism on risks and threats. We should establish an emergency response mechanism for AI safety and security incidents, formulate emergency plans, conduct emergency drills, and handle AI safety hazards, AI security threats, and events timely, rapidly, and effectively.
5.8 To enhance the training of AI safety talents.
We should promote the development of AI safety education in parallel with AI discipline. We should leverage schools and research institutions to strengthen talent cultivation in the fields of design, development, and governance for AIsafety. Support should be given to cultivating top AI safety talent in thecutting-edge and foundational fields, and also expanding such talent pool in areas such as autonomous driving, intelligent healthcare, brain-inspired intelligence and brain-computer interface.
5.9 To establish and improve the mechanisms for AI safety education, industry self-regulation, and social supervision.
We should strengthen education and training on the safe and proper use of AI among government, enterprises, and public service units. We should step up the promotion of knowledge related to AI risks and their prevention and response measures in order to increase public awareness of AI safety in all respects. We should guide and support industry associations in the fields of cybersecurity and AI to enhance industry self-regulation, and formulate self-regulation conventions that exceed regulatory requirements and serve exemplary roles. We should guide and encourage AI technology R&D institutions and service providers to continue to improve their safety capacity. A mechanism for handling public complaints and reports on AI risks and hazards should be established, forming an effective social supervision atmosphere for AI safety.
5.10 To promote international exchange and cooperation on AI safety governance.
We should actively make efforts to conduct cooperation with countries, support the building of an international
institution on AI governance within the United Nations framework to coordinate major issues related to AI development, safety, security, andgovernance. We should advance cooperation on AI safety governanceunder multilateral mechanisms such as APEC, G20 and BRICS, and strengthen cooperation with Belt and Road partner countries and Global South countries. Efforts should be made to study the matters relating to the construction ofan AI safety governance alliance to increase the representation and voice of developing countries in global AI governance. AI enterprises and institutions should be encouraged to engage in international exchanges and cooperation, share their best practices, jointly develop international standards of AI safety
