Safeguarding Culture in the Intelligent Media Era: Guided by Humanistic Values and Institutional Reconstruction in Multimodal AI Governance

Abstract

Multimodal AI technology is profoundly reshaping the cultural communication ecosystem. This study focuses on the intersection of multimodal AI and cultural security, proposing a dual-path governance framework centered on humanistic value-oriented intelligent assessment and innovation mechanisms. First, we construct a multi-modal data risk assessment model integrating text, images, and videos to enable dynamic perception and early warning of cultural security threats. Second, we design a four-dimensional innovation mechanism encompassing technology-coordination-institution-capacity. This framework aims to empower the modernization of cultural security governance, provide theoretical and technical solutions for mitigating systemic risks, and strategically serve national cultural security imperatives.