摘要：In terms of balancing AI application and governance, how should the international community build safe, reliable, and trustworthy

In terms of balancing AI application and governance, how should the international community build safe, reliable, and trustworthy AI systems? How can we enable all countries, especially developing ones, to fully participate in AI governance and achieve technological justice?

WANG Huiyao

President of Center for China and Globalization (CCG)

Building safe, reliable, and trustworthy artificial intelligence systems first requires the high attention of the international community. In March 2024, the United Nations General Assembly passed its first global resolution on artificial intelligence, emphasizing the need to develop safe, reliable, and trustworthy AI systems. The European Parliament also passed the Artificial Intelligence Act in March. In April, the United States and the United Kingdom signed a memorandum of understanding on artificial intelligence, announcing the establishment of a "Partnership for AI Safety and Scientific Cooperation". Earlier, on October 18, 2023, the Chinese government released the "Global Artificial Intelligence Governance Initiative" which systematically outlined China's approach to AI governance, focusing on development, security, and governance.

Second, safe, reliable, and trustworthy AI systems must ensure security, reliability, and explainability throughout their design and use. These systems must not only meet ethical and legal standards but also respect and protect human rights, and ensure privacy protection throughout their entire lifecycle, including design, development, testing, deployment, use, sale, procurement, operation, and decommissioning.

Third, it is essential to improve laws, regulations, and ethical guidelines, establish risk-level testing and evaluation systems, strengthen data security and privacy protection, enhance technical safety and reliability, and promote multi-stakeholder participation and cooperation.

In practice, developing countries face challenges in fully participating in global AI governance and international rule-making. This is partly due to the fact that many developing countries are still in the early stages of AI development, with limited awareness and motivation to engage in international rule-setting. Additionally, the dominance of developed countries in governance mechanisms and international rules further exacerbates the challenges. To ensure that all countries, especially developing ones, are fully involved in the AI governance process, it is crucial to enhance the representation and voice of developing countries in global AI governance. The principles of "equal rights", "equal opportunities", and "equal rules" should be upheld, along with international cooperation and assistance targeted at developing countries, to bridge the digital divide and governance capacity gap.

来源：澎湃新闻客户端