Delegation of Purchasing Tasks: The Role of AI Explainability and Anthropomorphism in Navigating Trust in the Digital Marketplace

How to cite

DENG, Zhixuan; LI, Chenyang; ZHANG, Xu. 2024. "Delegation of Purchasing Tasks: The Role of AI Explainability and Anthropomorphism in Navigating Trust in the Digital Marketplace." AEA RCT Registry. AEARCTR-0013784. https://doi.org/10.1257/rct.13784-1.3

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Zhixuan DENG* , Chenyang LI , Xu ZHANG

* Primary investigator

Registry ID
AEARCTR-0013784
DOI
https://doi.org/10.1257/rct.13784-1.3
First registered
Last updated
Status
On going
Trial dates
2024-01-19 – 2024-11-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/13784

Abstract

Based on behavioral economics theory and economic experimental methods, this study aims to explore whether consumers' perceived value, purchase decision, and trust in the entire online shopping process will be affected by AI explainability and anthropomorphism. By performing online and offline experiment, the study collect and analyze data on perceived capability, subjective trust and objective decision under different AI interaction conditions. This comprehensive approach is designed to ensure the broad applicability and far-reaching implications of the findings, which are expected to provide valuable insights into the application of AI to e-commerce platforms and lay a solid foundation for future research and practice in this field.

Experimental design

This study will adopt a 2*2 between-group experimental design. Participants were randomly assigned to one of four groups: control, XAI, Anthropomorphism, and XAI and Anthropomorphism. In the experimental group involving anthropomorphism, participants will be randomly assigned to either the male or female anthropomorphic group. After being assigned to different random groups, subjects will interact and play trust game with AI agent.

Design details

Randomization method
Randomization will be conducted by an online survey platform and the laboratory computer program.
Randomization unit
Participant (individual).
Clusters
1,000-1,500 adult subjects.
Observations
1,000-1,500 adult subjects.
Arms
300 subjects control, 300 subjects with XAI agent, 300 subjects with anthropomorphic AI agent, 300 subjects both treatment
Primary outcomes
Objective trust in AI, subjective trust in AI, AI ability perception, XAI degree, AI agent anthropomorphism degree

Cited by

No citation data available for this registration yet.

Keywords