Racial Bias in AI Training Data: Do Laypersons Notice?
Given that the nature of training data is the primary cause of algorithmic bias, do laypersons realize that systematic misrepresentation and under-representation of certain races in the training data can affect AI performance in a way that privileges some races over others? To answer this question,...
| Published in: | Media Psychology Vol. 29; no. 5; pp. 981 - 1009 |
|---|---|
| Main Authors: | , , |
| Format: | Article |
| Published: |
Taylor & Francis Ltd
2026
|
| Subjects: | |
| Online Access: | View this record in EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=196698591&site=ehost-live header: @attributes: shortDbName: ssf uiTerm: 196698591 longDbName: Social Sciences Full Text (H.W. Wilson) uiTag: AN controlInfo: bkinfo: jinfo: jid: 15213269 7MN jtl: Media Psychology issn: 15213269 maglogo: N pubinfo: dt: 2026 vid: 29 iid: 5 pid: 377 pub: Taylor & Francis Ltd artinfo: ui: 196698591 10.1080/15213269.2025.2558036 ppf: 981 ppct: 28 formats: tig: atl: Racial Bias in AI Training Data: Do Laypersons Notice? aug: au: Chen, Cheng Jang, Eunchae Sundar, S. Shyam affil: School of Communication, Oregon State University, Corvallis, OR, USA Media Effects Research Lab, Bellisario College of Communications, Pennsylvania State University, University Park, PA, USA Department of Immersive Media Engineering, Sungkyunkwan University, Seoul, Republic of Korea su: Algorithmic bias Emotion recognition Racism Public opinion Metadata Cognitive bias sug: subj: Algorithmic bias Emotion recognition Racism Public opinion Metadata Cognitive bias ab: Given that the nature of training data is the primary cause of algorithmic bias, do laypersons realize that systematic misrepresentation and under-representation of certain races in the training data can affect AI performance in a way that privileges some races over others? To answer this question, we conducted three between-subjects online experiments (N = 769 in total) with a prototype of an AI system that recognizes emotion-based facial expressions. Our results show that, by and large, training data representativeness is not an effective cue to communicate algorithmic bias. Instead, users rely on AI's performance bias to perceive racial bias in AI algorithms. In addition, the race of the users matters. Black participants perceive the system to be more biased when all facial images used to represent unhappy emotions in the training data are those of Black individuals. This finding highlights a significant human cognitive limitation that should be accounted for when communicating algorithmic bias arising from biases in the training data. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: N holdings: @attributes: islocal: N |
|---|