With the rapid development of media technology, the emergence of new communication technology makes the audience constantly facing novelty and unknown. As an intelligent media technology, artificial intelligence (AI) news anchor is a typical representative of "AI synthetic anchor", which is made by extracting the voice, lip shape, facial expressions and other features of the news broadcast videos of real anchors, combined with deep learning technology. The presence of a new robot or virtual image can lead people to ambiguous situations, and people may have feelings of unease that are difficult to describe or express. For example, when an virtual image or robot becomes very human-like but still behaves or looks slightly different from a human, people feel uneasy instead of increasing the perception of affinity. This phenomenon is called "uncanny valley" by Japanese robotics scientist Mori Masahiro.

This study designed three groups of experiments (N1=154, N2=104, N3=104), conducted nonparametric rank-sum test and linear regression analysis on the experimental results, aiming to prove the impact of AI news anchor's appearance, pronunciation and tone, and emotion on the audience's sense of uncanny. The results show that the AI news anchor will bring the Uncanny Valley effect in appearance anthropomorphism, that is, when the AI news anchors look like a human but can make the audience aware that they are not real people, they will bring the sense of uncanny; AI news anchors do not bring the Uncanny Valley effect in the aspects of pronunciation, tone and emotion anthropomorphism. The lower the degree of anthropomorphism in pronunciation, tone and emotion, the higher the audience's sense of uncanny. In the audience's perception, the factors causing uncanny are not everything related to ai news anchors. Appearance, voice and emotion are condensed in the real news anchor, but separated in the AI news anchors. Therefore, this study can better highlight the influence of various factors in inducing uncanny by controlling variables. When the audience is surrounded by mechanical electronic sound and emotionless expression, it is not surprising that the stereotype has taken root in their hearts. If the appearance of robots or real people is not considered when receiving news, the audience hopes to watch and listen the expression with emotion.

This research is practical to clarify the current situation and future of AI news anchors. It also explores the specific challenges and ethical issues brought about by intelligent technology to the traditional media industry. The emergence of AI news anchors is vital and essential, for example, replacing manual labor, having a lower error rate, and saving preparation time. However, the application of intelligent technology cannot solve all problems. At the same time, it might bring corresponding ethical issues, for example, triggering controversies about the rationality of the profession of real news anchors. But from the results of this study, it is clear that AI news anchors cannot replace real news anchors entirely, and may even trigger the Uncanny Valley effect.

News anchors do not convey information only through their voices, but rather as a whole of multiple elements to convey information. The standards for judging the talents in the news anchor industry are never the working hours, but the ability of innovation, expression and interpretation. The work of a real human anchor is a process of reprocessing and re-creating news content, or a process of artistic creation. The audience's perception of AI news anchors is also influenced by a variety of factors, and these factors have different effects on audience perception. The research in this paper confirms the importance of pronunciation, tone and emotion for news anchors. The more realistic the pronunciation, tone and emotion, the more it can dissipate the sense of uncanny that the robot may bring. This also reflects that not only the appearance of the news anchor is significant, but also the voice and emotion should not be neglected.