This paper develops an attitude-perception-intention (API) model of AI acceptance to explain individuals’ behavioral intention to accept AI-based recommendations as a function of attitude toward AI, trust and perceived accuracy with risk-level as a moderator. The API model was empirically validated through a between-participants experiment (N = 368) using a simulated AI-enabled investment recommendation system. One experimental condition depicted low-risk investment recommendation involving blue-chip stocks while the other depicted high-risk investment recommendation involving penny stocks. Attitude toward AI predicted behavioral intention to accept AI-based recommendations, trust in AI, and perceived accuracy of AI. Furthermore, risk level emerged as a significant moderator. When risk was low, a favourable attitude toward AI seemed sufficient to promote algorithmic reliance. However, when risk was high, a favourable attitude toward AI was a necessary but no longer sufficient condition for AI acceptance. The API model contributes to the human-AI interaction literature by not only shedding light on the underlying psychological mechanism of how users buy into AI-enabled advice but also adding to the scholarly understanding of AI recommendation systems in tasks that call for intuition in high involvement services such as finance where human counsel is usually preferred to machine-generated advice.
Bibliographical note© 2022 Elsevier Ltd. All rights reserved.This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy.
- AI-based recommendation
- Decision Sciences
- Investment decision
- Technology adoption