关键词: 人工智能生成内容, 可信性要素, 情景适配

Abstract: [Purpose/Significance] The trustworthiness of AI-Generated Content（AIGC）has emerged as a critical barrier to its broad deployment and social acceptance. A systematic review of relevant research can help clarify the foundational constructs underlying AIGC trustworthiness and provide theoretical support for future research on trustworthiness evaluation frameworks and governance models. [Design/Methodology] From the perspective of combining subjective perception and objective quality, this study dissected the concept of AIGC trustworthiness, clarifying its connotation and research scope. By adopting a systematic review approach, the multi-dimensional elements of AIGC trustworthiness were extracted and hierarchically analyzed. Then, it developed a conceptual framework of AIGC trustworthiness that encompassed four dimensions: users, content, environment, and technology. Furthermore, typical task scenarios were also categorized to explore the differences in key concerns and implementation approaches of AIGC trustworthiness elements. [Findings/Conclusion] This study identifies a set of multi-dimension and multi-layer AIGC trustworthiness elements, along with five categorized task scenarios, namely retrieval, execution, analysis, creation, and decision-making, each characterized by distinct trustworthiness concerns and strategic directions. Drawing on these insights, it outlines four future development pathways: methodological enhancement, quantitative evaluation, collaborative mechanisms, and scenario-based implementation. [Originality/Value] This study systematically elucidates the structural elements and scenario adaptability of AIGC trustworthiness. The results and recommendations offer theoretical insights to support the high-quality development of AIGC and its broader value realization.

Keywords: AI-Generated Content（AIGC）, Trustworthiness elements, Scenario adaptability