Paper Number
1525
Paper Type
Short Paper
Abstract
Predicting advice-taking is crucial for understanding user responses to artificial intelligence (AI) suggestions. Traditionally, advice-taking is measured as “weight of advice” (WOA) using pre- and post-advice evaluations, posing challenges in both online and real-world settings. To bridge the gap between the information systems and decision-making literature, we explore the predictive power of clickstreams as indicators of advice-taking, seeking to eliminate the need for explicit evaluations. In an online experiment, we examined the predictability of clickstreams in advice-taking when the credibility and the source of the advice are manipulated as independent variables. Our findings reveal significant relationships between source credibility, confidence, and advice-taking. Although advice from AI does not garner more acceptance, it prompts quicker pre-decisional reaction times, unveiling nuanced dynamics in the use of AI systems. This research provides insights into decision-making dynamics while proposing a novel approach for predicting advice-taking in AI systems.
Recommended Citation
Eizik, Meytal; Torgovitsky, Ilan; Haran, Uriel; and Fink, Lior, "Clickstreams as Traces of Weight of Advice" (2024). ECIS 2024 Proceedings. 3.
https://aisel.aisnet.org/ecis2024/track22_innrm/track22_innrm/3
Clickstreams as Traces of Weight of Advice
Predicting advice-taking is crucial for understanding user responses to artificial intelligence (AI) suggestions. Traditionally, advice-taking is measured as “weight of advice” (WOA) using pre- and post-advice evaluations, posing challenges in both online and real-world settings. To bridge the gap between the information systems and decision-making literature, we explore the predictive power of clickstreams as indicators of advice-taking, seeking to eliminate the need for explicit evaluations. In an online experiment, we examined the predictability of clickstreams in advice-taking when the credibility and the source of the advice are manipulated as independent variables. Our findings reveal significant relationships between source credibility, confidence, and advice-taking. Although advice from AI does not garner more acceptance, it prompts quicker pre-decisional reaction times, unveiling nuanced dynamics in the use of AI systems. This research provides insights into decision-making dynamics while proposing a novel approach for predicting advice-taking in AI systems.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.