視而信之?閱聽人對 AI 虛擬新聞主播與真人新聞主播可信度差異研究

Abstract

隨著生成式人工智慧技術快速發展，AI虛擬新聞主播走進閱聽人的視野，成為電視新聞中新興的傳播方式。然而，AI虛擬新聞主播是否能獲得與真人主播相同的信任與傳遞資訊的效果，仍是值得深入探討的議題。本研究旨在比較閱聽人對「真人電視新聞主播」 、 「AI虛擬新聞主播」的可信度與收視行為意圖差異，並進一步了解媒介使用習慣與人口變項對此關係的影響。

本研究以線上問卷實驗法進行，透過便利抽樣法在社群平台招募受測者，共取得506份有效樣本。受測者隨機觀看不同（真人/ AI）主播播報的國際新聞影片，並填寫可信度、收視意圖、媒介使用頻率量表，以分析不同主播的播報方式對閱聽人認知與收視行為的影響。

研究結論顯示，閱聽人對真人電視新聞主播的可信度與收視意圖均明顯高於AI虛擬新聞主播，不論 AI主播是否附加標籤，皆無法達到與真人主播相同的可信度、收視行為效果。整體而言，受測者仍較依賴真人主播的真實感、口語呈現與專業形象，並更願意持續觀看真人主播播報的內容。另一方面，媒介使用頻率雖與可信度評價有正相關，但並未明顯改變不同主播類型之間的可信度差異；人口變項部分，僅年齡對可信度具有顯著影響，年齡越高的閱聽人越傾向信任真人主播。

AI虛擬新聞主播雖具備技術創新與高效率等優勢，但在閱聽人信任的建立上仍有挑戰。未來新聞業若欲推動 AI主播的應用，需強化其語音自然度、表情真實感與肢體協調度，以逐步提升受眾的接受度與信任感。

With the rapid advancement of generative artificial intelligence technologies, AI virtual anchors have increasingly entered the public sphere, emerging as a new form of news delivery in television journalism. However, whether AI virtual news anchors can achieve the same level of trust and informational effectiveness as human news anchors remains an important question. This study aims to compare audience perceptions of credibility and viewing intentions toward human television news anchors and AI virtual news anchors, and to further examine how media usage habits and demographic variables influence this relationship.

An online experiment was conducted using a convenience sampling method, yielding 506 valid responses. Participants were randomly assigned to watch an international news video presented either by a human or an AI anchor and subsequently completed scales measuring credibility, viewing intentions, and media usage frequency. The study examines how different types of anchors influence audience cognition and viewing behavior.

The results indicate that audiences rate human news anchors significantly higher in both credibility and viewing intention compared to AI virtual news anchors. Regardless of whether the AI anchor was labeled, its perceived credibility and viewing effectiveness did not reach the level of a human anchor. Overall, participants relied more on the authenticity, natural speech delivery, and professional appearance of human anchors and were more willing to continue watching their content. Although media usage frequency showed a positive correlation with credibility evaluations, it did not significantly affect the credibility gap between different anchor types. Among demographic factors, only age had a significant effect—older viewers tended to place greater trust in human anchors.

While AI virtual news anchors provide benefits such as technological innovation and increased efficiency, building audience trust remains a significant challenge. For AI virtual news anchors to achieve broader adoption in the future, improvements in speech naturalness, facial expressiveness, and body coordination will be essential to enhancing audience acceptance and trust.