China issues guideline on ethics review for artificial intelligence

China has released a new guideline on the ethics review of artificial intelligence, setting out how AI systems should be assessed and managed to address potential risks.

The guideline, issued by multiple government departments including the Ministry of Industry and Information Technology, is presented as a trial framework. It outlines principles and processes for evaluating AI technologies before and during their deployment.

The document states that ethics reviews should focus on areas such as human well-being, fairness, and system reliability. It also highlights the importance of ensuring that AI systems remain controllable and trustworthy.

Specific aspects to be examined include the quality and selection of training data, the design of algorithms and models, and measures to reduce bias, discrimination, and misuse. The guideline also calls for technical tools to support risk assessment, auditing, and ongoing monitoring.

In addition, the framework encourages the development and sharing of datasets for ethics review purposes, as well as the promotion of AI systems that meet defined ethical standards. It also refers to the need to protect intellectual property in the development of such technologies.