Deepfake(深度造假)是Deep Learning(深度學習) + Fake(偽造)的混成詞,是任何意圖提供不實或錯誤資訊的「合成媒體」,包括文字、圖片、影像與聲音。而隨著網路傳播技術的進步與信息傳播的迅速,人工智慧技術亦被廣泛的應用在此領域,本研究就以探討Deepfake (深度偽造)議題為例,從同溫層效應去看Deepfake技術對假消息所帶來的衝擊與影響,與民眾認知不足的狀況,並嘗試提出一些解決方案與可行性。 本研究結果透過問卷調查可得知民眾對Deepfake技術的認知不足,而研究者為了要更瞭解民眾認知不足的原因和解決方案訪談了三位不同領域的專家,從專家訪談結論可得知,不論是新聞層面、技術層面、國安層面…等,都與民眾認知有著密不可分的關係,為了因應未來Deepfake所造成的影響,本研究得知提升民眾媒體識讀能力與Deepfake相關認知的重要,可比造美國媒體BuzzFeed製作一段假歐巴馬影片來宣導Deepfake技術的影響,最後再由相關單位澄清或解釋這是一部教育影片,以此來警戒民眾對於解讀訊息的正確態度以及使用第三方查核機制、相關辨識工具的重要性。我們無法阻擋AI科技的發展,當Deepfake在台灣普及化之後,民眾若是建立起好的媒體識讀能力,多少能防範這種類型的假消息。
With the advancement of Internet communication technology and the rapid information dissemination, artificial intelligence technology has also been widely used in this field. This study will take the discussion of Deepfake as an example, and look at Deepfake technology from the stratospheric effect. In response to the impact and influence of false news, and the insufficient public awareness, we will try to propose some solutions and feasibility. The results of this study can be learned through questionnaire surveys that the public’s awareness of Deepfake technology is insufficient, and the researchers interviewed three experts in different fields in order to better understand the causes and solutions of the public’s lack of cognition. It is the news level, the technical level, the national security level, etc., which are inseparably related to the public perception. In order to cope with the future impact of Deepfake, this research has learned that the importance of improving the public’s media literacy and deepfake-related cognition is comparable. We can't stop the development of AI technology. After the popularization of Deepfake in Taiwan, if the people build up good media literacy skills, they can more or less prevent this type of false news.