Explanation strategies in humans versus current explainable artificial intelligence: Insights from image classification.
Explainable AI (XAI) methods provide explanations of AI models, but our understanding of how they compare with human explanations remains limited. Here, we examined human participants' attention strategies when classifying images and when explaining how they classified the images through eye‐trackin...
| Publicado en: | British Journal of Psychology Vol. 117; no. 2; pp. 479 - 503 |
|---|---|
| Autores principales: | , , , , |
| Formato: | Artículo |
| Publicado: |
Wiley-Blackwell
May2026
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=192785876&site=ehost-live header: @attributes: shortDbName: ssf uiTerm: 192785876 longDbName: Social Sciences Full Text (H.W. Wilson) uiTag: AN controlInfo: bkinfo: jinfo: jid: 00071269 BJP jtl: British Journal of Psychology issn: 00071269 maglogo: Y pubinfo: dt: May2026 vid: 117 iid: 2 pid: 480 pub: Wiley-Blackwell artinfo: ui: 192785876 10.1111/bjop.12714 ppf: 479 ppct: 24 formats: fmt: – @attributes: type: T – @attributes: type: C – @attributes: type: P size: 1.6MB tig: atl: Explanation strategies in humans versus current explainable artificial intelligence: Insights from image classification. aug: au: Qi, Ruoxi Zheng, Yueyuan Yang, Yi Cao, Caleb Chen Hsiao, Janet H. affil: Department of Psychology, University of Hong Kong, Hong Kong SAR, China Huawei Research Hong Kong, Hong Kong SAR, China Big Data Institute, Hong Kong University of Science and Technology, Hong Kong SAR, China Division of Social Science, Hong Kong University of Science and Technology, Hong Kong SAR, China su: China Task performance Artificial intelligence Cell phones Attention Analysis of variance Research funding Corn Winter sports Portable computers T-test (Statistics) Tennis Multiple regression analysis Eye movement measurements Lemon Analysis of covariance Large-scale brain networks Concepts Visual perception Insects Mollusks Horses Mushrooms Transducers sug: subj: Task performance Artificial intelligence Cell phones Attention Analysis of variance China Postharvest Crop Activities (except Cotton Ginning) Corn Farming Other Farm Product Raw Material Merchant Wholesalers Live animal merchant wholesalers Citrus (except Orange) Groves Shellfish Farming Mushroom Production Electronic Computer Manufacturing Computer and peripheral equipment manufacturing Other Electronic Component Manufacturing Semiconductor and other electronic component manufacturing Electronic components, navigational and communications equipment and supplies merchant wholesalers Wireless Telecommunications Carriers (except Satellite) Radio and Television Broadcasting and Wireless Communications Equipment Manufacturing Electronics Stores Research funding Corn Winter sports Portable computers T-test (Statistics) Tennis Multiple regression analysis Eye movement measurements Lemon Analysis of covariance Large-scale brain networks Concepts Visual perception Insects Mollusks Horses Mushrooms Transducers keyword: EMHMM explainable AI explanation eye movements image classification text analysis EMHMM explainable AI explanation eye movements image classification text analysis ab: Explainable AI (XAI) methods provide explanations of AI models, but our understanding of how they compare with human explanations remains limited. Here, we examined human participants' attention strategies when classifying images and when explaining how they classified the images through eye‐tracking and compared their attention strategies with saliency‐based explanations from current XAI methods. We found that humans adopted more explorative attention strategies for the explanation task than the classification task itself. Two representative explanation strategies were identified through clustering: One involved focused visual scanning on foreground objects with more conceptual explanations, which contained more specific information for inferring class labels, whereas the other involved explorative scanning with more visual explanations, which were rated higher in effectiveness for early category learning. Interestingly, XAI saliency map explanations had the highest similarity to the explorative attention strategy in humans, and explanations highlighting discriminative features from invoking observable causality through perturbation had higher similarity to human strategies than those highlighting internal features associated with higher class score. Thus, humans use both visual and conceptual information during explanation, which serve different purposes, and XAI methods that highlight features informing observable causality match better with human explanations, potentially more accessible to users. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: N holdings: @attributes: islocal: N |
|---|