Building a Safe Framework for Artificial Intelligence

Recently, the Central Political Bureau meeting emphasized the need to “improve artificial intelligence governance” while deploying the development of a new form of intelligent economy. This reflects a balanced approach between development and regulation, which will further promote the healthy and orderly development of artificial intelligence.

As a strategic technology leading a new round of technological revolution and industrial transformation, artificial intelligence is profoundly changing human production and lifestyle, becoming an important force in reshaping the global competitive landscape and promoting economic and social transformation. Currently, China is fully implementing the “AI+” initiative, with breakthroughs in large models, intelligent agents, and embodied intelligence technologies, leading to the rapid realization of numerous high-efficiency application scenarios. However, technological development also brings multiple risks such as algorithmic bias, data misuse, and deepfakes. In the face of the rapid evolution of AI technology, its extensive applications, and the complex intertwining of risks, strengthening governance to promote AI development in a beneficial, safe, and equitable direction is a significant challenge for society.

Risk prevention and regulation are core tasks of governance. AI technology has a “black box” nature, with opaque decision-making logic and difficult-to-trace outputs, posing unknown safety hazards. Therefore, it is necessary to explore new regulatory concepts and methods, providing a testable and dynamically inclusive governance environment for technological iteration, allowing enough room for innovation while maintaining safety.

Currently, many regions in China are actively exploring new regulatory methods such as sandbox regulation, allowing relevant entities to pilot new products, services, models, and technologies within limited scenarios and scales. At the same time, inclusive and prudent regulatory measures are adopted to allow for error correction within controllable limits, preventing risk spillover, and achieving positive results. In the future, it will be necessary to accelerate the improvement of relevant institutional arrangements, clarifying the access, operation, and exit rules for sandbox regulation. For instance, adhering to a classification and grading principle, stricter regulatory standards and operational processes should be implemented for high-risk areas involving data security and financial security, while conditions may be moderately relaxed in areas with mature technologies and lower potential spillover risks. Additionally, continuously optimizing regulatory technology tools will enhance regulatory capabilities in line with AI technology, achieving dynamic regulation throughout the entire lifecycle.

Improving the legal system is the fundamental guarantee of governance. In recent years, China has continuously advanced institutional construction in key areas such as generative AI, algorithm recommendations, and deep synthesis, issuing a series of regulatory rules like the “Interim Measures for the Management of Generative AI Services” and the “Regulations on Algorithm Recommendation Management for Internet Information Services,” accumulating rich governance experience. However, during the current and forthcoming explosive development phase of AI technology, some regulations are lagging behind technological innovations. It is essential to deeply understand the trends and laws of AI development, expedite the improvement of relevant laws, regulations, policies, application norms, and ethical guidelines, and timely delineate the ethical boundaries and safety bottom lines of technology applications, ensuring that legal development keeps pace with technological advancements.

Strengthening international cooperation is an inevitable choice for governance. AI is a universal technology that transcends national borders, and its risks are global in nature; thus, normative governance is not solely a matter for one country or region. However, due to the high sensitivity of technological competition, there are still differences among countries regarding cross-border data, content review, and government regulatory authority. It is necessary to promote the alignment and coordination of development strategies, governance rules, and technical standards among all parties, aiming to establish a globally recognized framework and standards for AI governance as soon as possible.

As a responsible major country, China has actively participated in and led global AI governance in recent years, successively releasing important documents such as the “Global AI Governance Initiative” and the “Global AI Governance Action Plan.” In the future, China should continue to promote innovation in AI governance systems and practices, contributing more Chinese solutions to the international community, and strengthen international exchanges and cooperation to build a fairer and more reasonable global governance system.

AI governance is a systematic integration of technical ethics, legal construction, and international cooperation. A multi-faceted approach is essential to build a solid safety barrier. By adhering to the dual focus on development and safety, as well as innovation and regulation, exploring and improving the AI governance system will provide a more robust institutional guarantee for the development of a new form of intelligent economy.