Suppr超能文献

语音声学如何影响说话者的可信度感知?一项系统综述。

How do voice acoustics affect the perceived trustworthiness of a speaker? A systematic review.

作者信息

Maltezou-Papastylianou Constantina, Scherer Reinhold, Paulmann Silke

机构信息

Department of Psychology and Centre for Brain Science, University of Essex, Colchester, United Kingdom.

Brain-Computer Interfaces and Neural Engineering Laboratory, School of Computer Science and Electronic Engineering, University of Essex, Colchester, United Kingdom.

出版信息

Front Psychol. 2025 Mar 10;16:1495456. doi: 10.3389/fpsyg.2025.1495456. eCollection 2025.

Abstract

Trust is a multidimensional and dynamic social and cognitive construct, considered the glue of society. Gauging someone's perceived trustworthiness is essential for forming and maintaining healthy relationships across various domains. Humans have become adept at inferring such traits from speech for survival and sustainability. This skill has extended to the technological space, giving rise to humanlike voice technologies. The inclination to assign personality traits to these technologies suggests that machines may be processed along similar social and vocal dimensions as human voices. Given the increasing prevalence of voice technology in everyday tasks, this systematic review examines the factors in the psychology of voice acoustics that influence listeners' trustworthiness perception of speakers, be they human or machine. Overall, this systematic review has revealed that voice acoustics impact perceptions of trustworthiness in both humans and machines. Specifically, combining multiple acoustic features through multivariate methods enhances interpretability and yields more balanced findings compared to univariate approaches. Focusing solely on isolated features like pitch often yields inconclusive results when viewed collectively across studies without considering other factors. Crucially, situational, or contextual factors should be utilised for enhanced interpretation as they tend to offer more balanced findings across studies. Moreover, this review has highlighted the significance of cross-examining speaker-listener demographic diversity, such as ethnicity and age groups; yet, the scarcity of such efforts accentuates the need for increased attention in this area. Lastly, future work should involve listeners' own trust predispositions and personality traits with ratings of trustworthiness perceptions.

摘要

信任是一个多维度、动态的社会和认知概念,被视为社会的黏合剂。判断一个人是否值得信任对于在各个领域建立和维持健康的人际关系至关重要。人类已经善于从言语中推断这些特质以实现生存和可持续发展。这项技能已经延伸到技术领域,催生了类人语音技术。给这些技术赋予个性特质的倾向表明,机器可能会像人类声音一样沿着相似的社会和声音维度被处理。鉴于语音技术在日常任务中的日益普及,本系统综述考察了语音声学心理学中影响听众对说话者(无论是人类还是机器)可信度感知的因素。总体而言,本系统综述表明,语音声学会影响对人类和机器的可信度感知。具体而言,与单变量方法相比,通过多变量方法结合多种声学特征可提高可解释性,并产生更平衡的结果。在不考虑其他因素的情况下,单独关注音高之类的孤立特征在综合各项研究来看时往往会得出不确定的结果。至关重要的是,情境或背景因素应被用于加强解释,因为它们往往能在各项研究中提供更平衡的结果。此外,本综述强调了交叉检验说话者-听众人口统计学多样性(如种族和年龄组)的重要性;然而,此类研究的匮乏凸显了在这一领域加大关注力度的必要性。最后,未来的研究工作应涉及听众自身的信任倾向和个性特质以及可信度感知评分。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/d317/11931160/90aaedd7f235/fpsyg-16-1495456-g001.jpg

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验