US and China need shared risk vocabulary to manage AI threats: Beijing-linked think tank

The central idea from the CICIR is that effective management of AI threats requires a common understanding of terminology before addressing broader technological competition. This approach suggests that establishing a shared lexicon for AI risks could build a foundation of trust, enabling more productive bilateral discussions on AI governance and incident response. The next dialogue is scheduled for November.

The proposed dialogue mechanism, agreed upon by both presidents, aims to facilitate exchanges on AI's risks and benefits, and to establish a bilateral communication channel for AI-related incidents. This structure points to a focus on operational aspects of AI management, rather than immediate policy harmonization. Liu Chong, director of CICIR's Centre for Sci-Tech and Strategic Studies, stated that the dialogue should begin with this shared understanding to gradually build trust.

CICIR's emphasis on a 'dynamic system of risk changes and response levels' instead of 'abstract red lines' indicates a preference for adaptive frameworks over rigid boundaries. This suggests a recognition that AI risks are evolving and require flexible response strategies, which necessitates a common descriptive language to even identify and discuss these changes effectively.