Transparency and Public Communication Foster Trust in AI Companies

This study examines how organizational characteristics of companies producing Artificial Intelligence (AI) technologies influence public trust through a vignette-based experimental design. Building on prior frameworks of trust, we focus on five sub-dimensions of trust: Benevolence, Standards and Guidelines, Data Quality, Reliability, and Transparency, each with three different levels. Results indicate that Transparency and Benevolence are the most significant drivers of trust. Organizations that provide clear explanations of their AI technologies and demonstrate societal accountability by seeking and incorporating public feedback are viewed more favorably. Adherence to external standards, such as national or international guidelines, further enhances trust, while technical performance and data quality are less influential, as participants assume the technology is functioning adequately for their limited use. We conclude that transparent practices, societal engagement, and institutional collaboration will foster public confidence in companies producing AI technologies.

本研究采用基于情景实验（vignette）的实验范式，探究人工智能（Artificial Intelligence, AI）技术研发企业的组织特征对公众信任的影响作用。本研究依托既有信任研究框架，聚焦信任的五大子维度：善意（Benevolence）、标准与规范（Standards and Guidelines）、数据质量（Data Quality）、可靠性（Reliability）以及透明度（Transparency），每个维度均设置三个不同水平。研究结果表明，透明度与善意是影响公众信任的最显著驱动因素。能够对其AI技术作出清晰阐释，并通过征集并吸纳公众反馈以彰显社会责任感的企业，将获得公众更高的认可度。遵循国家或国际指南等外部标准的行为，可进一步强化公众信任；而技术性能与数据质量的影响则相对有限，这是由于实验被试认为，在其有限的使用场景中，该技术已能够满足基本运行需求。本研究最终得出结论：透明化实践、社会参与以及制度层面的协作，将有助于培育公众对人工智能技术研发企业的信任信心。