The Psychological Acceptability of AI-Generated Business Communication: A Quantitative Study of Stakeholder Perceptions and Attribution Effects
Artificial Intelligence (AI) is increasingly used in business communication, yet stakeholder responses to AI-generated language remain insufficiently understood. This study investigates the psychological acceptability of AI-generated business communication by examining how disclosure of authorship influences stakeholder perceptions. Drawing on Uncanny Valley Theory (Mori, 1970) and Attribution Theory (Weiner, 1985), the research explores the psycholinguistic mechanisms that shape evaluations of AI-generated messages. A quantitative experimental design was employed involving 100 stakeholders, including employees, customers, and investors, who assessed business communications presented under three conditions: AI-generated and disclosed as AI-written, human-written, and AI-generated without disclosure. The findings reveal a significant AI attribution penalty. Identical texts received substantially lower authenticity ratings when labeled as AI-generated, with disclosed AI messages rated 24% less authentic than undisclosed AI texts (t = 3.86, p < .001). Participants identified AI authorship correctly only 52% of the time, indicating near-chance detection ability. While AI-generated communication was perceived as clearer than human-written text (M = 3.94 vs. 3.62), it scored lower on warmth (M = 2.86 vs. 3.72) and personality (M = 2.68 vs. 3.84). Acceptability also varied across communication contexts, reaching its highest level in informational reports (78% acceptable) and its lowest in empathetic messages (34% acceptable). Multiple regression analysis showed that perceived authenticity (β = .38), communication context sensitivity (β = .28), and disclosure transparency (β = .22) jointly explained 44% of the variance in stakeholder acceptability. The study contributes to linguistic, psychological, and management scholarship by illuminating factors that influence stakeholder trust and acceptance of AI-mediated business communication.
Paper
The full text of this publication is not hosted on 44B due to licensing.
Read it at OpenAlex