- China released new AI ethics review rules aimed at regulating industry development and preventing potential social risks brought by new technologies.
- The new regulations establish a strict expert review mechanism for high-risk AI activities.
China has released measures for AI science and technology ethics review and services, marking a significant step in the country's AI regulation and governance.
China's Ministry of Industry and Information Technology (MIIT) and nine other departments jointly issued the trial provisions, aiming to balance technological innovation with risk prevention and ensure the high-quality development of the AI industry.
The new rules clarify the scope and procedures for the ethical review of AI scientific and technological activities, focusing on six core aspects including human well-being, fairness and justice, controllability and trustworthiness, and privacy protection, according to a document published by the MIIT on Tuesday.
Enterprises, universities, and research institutions engaged in AI research and development, acting as the responsible entities, are required to establish dedicated ethics committees to independently conduct sci-tech ethics reviews.
Tailored to different risk levels, the measures formulate differentiated review procedures — including general, simplified, expert review, and emergency procedures — to standardize the ethical governance of AI activities.
For the research and development of algorithmic models with public opinion and social mobilization capabilities, or highly autonomous automated decision-making systems, the activities must undergo expert reviews organized by authorities.
While strengthening oversight, the new rules also outline several supportive measures. If an AI activity has already fulfilled regulatory approval in relevant areas, the expert review process can be directly exempted, the provisions said.
The Chinese government will also promote innovation in ethics review technologies and the open-sourcing of high-quality datasets, effectively helping micro, small, and medium-sized enterprises enhance their capacity to prevent and control AI ethics risks.
