Original
The impact of public agency representation on citizen perception in the face of technological discrimination
Abstract
This study explores how public agency representation and expertise shape citizen perceptions of algorithmic discrimination. Using theories of representative bureaucracy and Human-In-The-Loop (HITL), an experimental design with a general population and an African American sample examines how algorithmic correction approaches influence trust, legitimacy, and perceived performance. Expert involvement increases positive perceptions across samples, with stronger effects observed in the African American sample. Awareness of inclusion boosts trust among the African American sample but has mixed results for the general population. Findings suggest that while technical expertise is crucial, representative inclusion is vital for rebuilding trust within communities historically impacted by discrimination.
中文
公共机构代表性对技术歧视情境下公民感知的影响
摘要
本研究探讨公共机构代表性与专业知识如何塑造公民对算法歧视的感知。基于代表性官僚制与人在回路（Human-In-The-Loop, HITL）理论，本研究采用面向普通公众与非洲裔美国人样本的实验设计，考察算法纠偏方式如何影响信任、合法性与绩效感知。专家参与提升了两个样本中的正面感知，且在非洲裔美国人样本中效应更强。对包容性的知晓提升了非洲裔美国人样本的信任，但在普通公众中结果不一。研究发现表明，技术专业知识固然关键，但代表性包容对于在历史上受歧视影响的社群中重建信任至关重要。
关键词
代表性官僚制、算法歧视、公民感知、信任、合法性、人在回路、实验研究、公共机构代表性