AI深度偽造影像的社會論述與真實性想像,ERICDATA高等教育知識庫
高等教育出版
熱門: Scopus AI  Semi-trusted loader  吳清山  Helong Xia  Qiaoyun Zhang  Zhi-Wei Sun  
高等教育出版
首頁 臺灣期刊 學校系所 學協會 民間出版 大陸/海外期刊 政府機關 學校系所 學協會 民間出版 DOI查詢服務 DOI註冊服務
篇名
AI深度偽造影像的社會論述與真實性想像
並列篇名
The Social Discourses and Imaginaries of Authenticity in AI Deepfake Images
作者 張重金
中文摘要
人工智慧生成的深度偽造影像挑戰「眼見為憑」,引發真實性與倫理爭議。既有研究多聚焦於技術與法律,卻忽略觀眾的日常實踐。本研究以臺灣青年為核心,蒐集TikTok與YouTube留言,並結合深度訪談,運用平台留言之質性內容分析(QCA;兼具論論述分析視角)與詮釋現象學分析;(IPA),探討真假協商與倫理想像。研究發現:觀眾反應涵蓋驚訝、懷疑、娛樂與冷漠,真假判斷呈現「懷疑—尋證—協商」的動態歷程兼倫理討論則聚焦同意、風險與治理,顯示焦慮與娛樂化並存。臺灣青年展現「靈活的懷疑」與「娛樂化解讀」,有別於西方研究所凸顯的嚴肅焦慮。本研究揭示真實性乃語言化、文化化的動態建構,並對媒體素養教育與政策提出在地啟示。
英文摘要
AI-generated deepfake images challenge the assumption that“seeing is believing,”raising questions about authenticity and ethics. Existing scholarship has largely focused on technological and legal dimensions, while paying less attention to audiences’everyday interpretive practices. Focusing on Taiwanese youth, this study draws on comments collected from TikTok and YouTube together with in-depth interviews. Platform comments were analyzed using qualitative content analysis (QCA) with a discourse-analytic perspective, while interview data were examined through interpretative phenomenological analysis (IPA) to explore negotiations of authenticity and ethical imaginaries. The findings show that audience responses range from surprise and skepticism to amusement and indifference. Judgments of authenticity unfold through a dynamic process of“suspicion–evidence seeking–negotiation,”while ethical discussions center on consent, risk, and governance, revealing the coexistence of anxiety and entertainment-oriented framings. Taiwanese youth display“flexible skepticism”and“entertainment-oriented interpretation,”contrasting with the stronger emphasis on anxiety found in much of the Western scholarship reviewed in this study. This study shows that authenticity is dynamically constructed through language and culture and offers locally grounded implications for media literacy education and policy development.
起訖頁 1-41
關鍵詞 人工智慧、真實性、倫理評價、深度偽造、媒體素養、Artificial Intelligence、Authenticity、Ethical Evaluation、Deepfake、Media Literacy
刊名 人文社會科學研究  
期數 202606 (20:2期)
出版單位 國立屏東科技大學人文暨社會科學院
該期刊
下一篇
非人類記憶的轉向:當AI開始書寫自己的藝術文化史

高等教育知識庫  新書優惠  教育研究月刊  Ericdata高等教育知識庫