Abstract: The risks posed by artificial intimacy have become an important issue in global AI governance. Taking“AI anthropomorphic interaction” as the core concept, this paper clarifies the generation mechanism and specific risks of artificial intimacy. Based on the framework of “Trustworthy AI”, this paper refines governance principles of artificial intimacy and constructs an analytical framework. Based on this, this paper examines recent laws and legislative drafts for public comments issued by the Cyberspace Administration of China, and New York State and California of the United States, thereby revealing divergences in governance paths between the two countries. The findings indicate that U.S. governance tends to be “negative governance based on individual rational choice”, which treats AI products as a value-neutral tool, leaving the scale and trajectory of artificial intimacy to individual autonomy while restricting legal intervention exclusively to extreme cases such as suicide and self-harm. Conversely, China’s governance tends to be “active regulation embedded in real social relations and values, which acknowledges user vulnerability in artificial intimacy, emphasizes the integration of ethical values into AI design and further incorporates interactional risks into social network and value system to provide comprehensive institutional safeguards through institutional forces.
Keywords: Artificial Intimacy; AI Anthropomorphic Interaction; Trustworthy AI; AI Governance
Author: XIN Yanyan
Source: Xin Y. Y. (2026). Value Embedding and Procedural Safeguards: A Comparative Study of China and the United States on the Governance of Artificial Intimacy. Journal of CUPL, (3).
Publish date：2026/5/12
Original Link:
