Paper Type
ERF
Abstract
As AI becomes embedded in managerial decision-making, understanding how to configure AI roles to augment human capabilities is critical. While current approaches predominantly configure AI as a recommender, we propose that AI can be alternatively configured as a Devil's Advocate (Challenger-AI), questioning assumptions and surfacing potential weaknesses. Drawing on Cognitive Theories, we develop a dual-pathway model: A Challenger-AI enhances decision outcomes through increased information elaboration while imposing additional mental workload. Our study tests main effects, mediators, and control variables by comparing the results of using a Recommender-AI versus a Challenger-AI on decision quality and confidence, examining dual-pathway mechanisms and including individual control variables. This research contributes to human-AI collaboration literature by examining whether AI configured to challenge rather than recommend can improve complex managerial decision-making.
Paper Number
1298
Recommended Citation
Liu, Shuqing; Manson, Kerr; Galletta, Dennis; and Ware, Thomas, "Beyond the Recommender: How AI Role Configurations Shape Complex Managerial Decisions" (2026). AMCIS 2026 Proceedings. 8.
https://aisel.aisnet.org/amcis2026/sig_dsa/sig_dsa/8
Beyond the Recommender: How AI Role Configurations Shape Complex Managerial Decisions
As AI becomes embedded in managerial decision-making, understanding how to configure AI roles to augment human capabilities is critical. While current approaches predominantly configure AI as a recommender, we propose that AI can be alternatively configured as a Devil's Advocate (Challenger-AI), questioning assumptions and surfacing potential weaknesses. Drawing on Cognitive Theories, we develop a dual-pathway model: A Challenger-AI enhances decision outcomes through increased information elaboration while imposing additional mental workload. Our study tests main effects, mediators, and control variables by comparing the results of using a Recommender-AI versus a Challenger-AI on decision quality and confidence, examining dual-pathway mechanisms and including individual control variables. This research contributes to human-AI collaboration literature by examining whether AI configured to challenge rather than recommend can improve complex managerial decision-making.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIG DSA