On October 2, reports surfaced indicating that the prospects for China-US cooperation in the realm of artificial intelligence (AI) appear more promising than they have been in recent months. Sources privy to the discussions within the Chinese negotiation team disclosed that both nations intend to forge ahead with specific collaborative initiatives during their upcoming dialogue in November, focusing on at least three pragmatic avenues.

Firstly, the two countries aim to collaboratively devise classification and grading criteria for AI risks. This endeavor involves delineating major risk scenarios and instances of severe model失控 (loss of control in Chinese, here meaning the uncontrollable situations of AI models), thereby laying a solid foundation for subsequent deliberations on intricate regulatory frameworks.

Secondly, they will delineate the scope, procedures, and requisites for reporting AI risk incidents. This includes establishing effective communication channels and specifying the types of incidents to be reported, along with the deadlines for reporting, the content of information to be shared, and the measures to be taken in cases of non-compliance with reporting obligations.

Thirdly, the integration of AI governance into the broader framework of China-US exchanges will be pursued. This move seeks to ensure comprehensive participation from the Chinese side and amplify the voices of a wider array of countries in the global discourse on AI governance.