 |
| October 2026 |
 |
78 |
|
| 專輯論文Special Issue Articles |
| 社交媒體環境下公眾對生成式AI信任的形成機制研究 |
| Exploring Public Trust in Generative AI and Its Formation Mechanism on Social Media |
|
|
|
 |
(9)
|
| 作者 |
高芳芳、徐欣妍 |
| Author |
Fangfang GAO, Xinyan XU |
| 關鍵詞 |
生成式AI信任、社交媒體接觸、AI知識、AI效能感、AI風險感知 |
| Keywords |
trust in generative AI, social media exposure, AI knowledge, AI efficacy, AI risk perception |
| 摘要 |
信任是影響生成式AI社會應用與推廣的關鍵因素。本研究基於問卷調查（N = 787），採用結構方程模型（structural equation modeling, SEM）與模糊集定性比較分析（fuzzy-set qualitative comparative analysis, fsQCA）相結合的混合方法，考察了社交媒體環境下公眾對生成式AI信任的形成機制。 SEM的結果表明，社交媒體接觸能通過強化AI知識、增強AI效能感及降低AI隱私風險感知三條路徑來促進公眾對生成式AI的技術信任與系統信任。但社交媒體接觸僅能通過降低社會風險感知來提高技術信任，卻無法經由這一路徑對系統信任產生顯著影響。fsQCA分析進一步揭示了上述因素對生成式AI信任的組態效應。系統信任的形成組態包括效能驅動－風險規避型、知識驅動－風險規避型與社交媒體賦能型三類。技術信任的形成組態更為多元，在上述三類組態的基礎上還涵蓋媒介驅動－風險規避型。本研究拓展了對生成式AI信任的概念框架，釐清了技術信任與系統信任的差異化形成機制，並為政策制定者和技術開發者提供了實踐建議。 |
| Abstract |
Trust is a key factor influencing the social adoption and diffusion of generative AI. Drawing on a survey (N = 787), this study employed a mixed-methods approach that combines structural equation modeling (SEM) and fuzzy-set qualitative comparative analysis (fsQCA) to examine the mechanisms underlying public trust in generative AI in social media contexts. The SEM results indicate that social media exposure can promote both technical trust and system trust in generative AI by enhancing AI knowledge, strengthening AI efficacy, and reducing perceived privacy risks. However, social media exposure can only improve technical trust by reducing perceived social risks, while this pathway has no significant effect on system trust. The fsQCA analysis further reveals the configurational effects of these factors on trust in generative AI. The formation of system trust in generative AI is characterized by three configurations, including efficacy-driven/risk-averse, knowledge-driven/risk-averse, and social media-empowered. The configurations for technical trust are more diverse, additionally incorporating a media-driven/risk-averse configuration beyond the three configurations identified for system trust. This study expands the conceptual framework of trust in generative AI, clarifies the different formation mechanisms of technical trust and system trust, and offers practical suggestions for policymakers and AI developers.
本文引用格式：
高芳芳、徐欣妍（2026）。〈社交媒體環境下公眾對生成式AI信任的形成機制研究〉。《傳播與社會學刊》，第78期，頁111–142。
Citation of This Article:
Gao, F., & Xu, X. (2026). Exploring public trust in generative AI and its formation mechanism on social media. Communication and Society, 78, 111–142. |
|
|
 |
| No.77 2026 July |
 |
| No.76 2026 April |
 |
| No.75 2026 January |
 |
| No.74 2025 October |
 |
| No.73 2025 July |
|
