Economic Daily: Establishing Policies and Systems to Support the Artificial Intelligence Industry

Data shows that by 2025, the number of artificial intelligence companies in China will exceed 6,000, with the core industry scale expected to surpass 1.2 trillion yuan. Currently, AI applications have covered key industries such as steel, non-ferrous metals, electricity, and telecommunications, gradually penetrating critical links in product research and development, quality inspection, and customer service. As a technology leading a new wave of scientific and technological revolution, artificial intelligence is profoundly reshaping economic structures and social governance models. Accelerating the construction of a legal framework compatible with AI is a key measure to ensure its steady and long-term development.

From a global perspective, AI governance presents diverse approaches. The European Union’s Artificial Intelligence Act establishes a risk-based regulatory model, creating a four-tier supervision system of prohibited, high-risk, limited-risk, and minimal-risk categories. The United States adopts an innovation-oriented regulatory strategy, emphasizing standardization and industry self-discipline to promote innovation through the AI Executive Order. China’s governance approach is characterized by the concept of “development and security in tandem, innovation and regulation in coordination,” exploring mechanisms for data rights allocation and circulation that safeguard data security while facilitating data market circulation, providing valuable reference for global AI governance.

It is also important to recognize that China’s AI legal framework still faces many challenges. Legally, there is a lack of specialized laws, and coordination among laws such as the Cybersecurity Law and Data Security Law is not yet perfect. In regulatory implementation, issues such as unclear departmental responsibilities, overlapping functions, and inconsistent standards persist, and the balance between algorithm transparency requirements and trade secret protection needs improvement. In technical governance, problems such as uneven data quality, difficulty eliminating algorithmic bias, and vague responsibility boundaries remain unresolved. Additionally, there are lagging regulations in areas like intellectual property protection and cross-border data flows. Looking ahead to the 14th Five-Year Plan, multi-faceted measures are needed to establish systems that support the development of the AI industry.

In legislative development, efforts should focus on building a regulatory system centered on graded and classified supervision supported by a comprehensive technical standards system. For regulatory innovation, establishing cross-departmental collaborative supervision platforms, unifying enforcement standards, and implementing “regulatory sandbox” mechanisms in specific fields such as autonomous driving, along with setting up innovation pilot zones, will promote innovation while ensuring safety.

In data governance, breakthroughs are needed in defining property rights, potentially through comprehensive systems covering data ownership, processing and usage rights, and data product operation rights. Establishing management mechanisms that cover the entire data lifecycle—from collection and use to destruction—is essential, especially in standardizing training data quality assessment and data annotation norms.

In algorithm accountability, a responsibility system covering the entire process from design and development to deployment should be established, with mandatory evaluation requirements in high-risk areas. Introducing algorithm impact assessments, requiring developers to evaluate fairness, transparency, and security before deployment, and granting users the right to explanations and objections are also necessary.

Furthermore, strengthening corporate responsibility, integrating ethical review into R&D processes, and encouraging the establishment of algorithm governance committees are important. Industry associations should be guided to formulate higher-level industry standards and technical norms, and an AI ethics certification system should be established. Improving public participation mechanisms through hearings, expert evaluations, and opinion surveys can help build social consensus and ensure that technological development aligns with public interests. Active participation in global rule-making is essential to promote the establishment of a fair, inclusive, and sustainable global AI governance order.

(Source: Economic Daily)