Revisiting Warranting Theory in AI-Mediated Communication: An Evaluation of Writer and Source Effects in Hotel Reviews

This paper evaluates how AI-Mediated Communication (AIMC) influences impression formation. By drawing on and extending warranting theory, this paper examined how source roles (self vs. third party) and writer labels (AI vs. human) interact to affect impression formation and message effects in the AIMC setting of hotel reviews. Two pre-registered experiments (Study 1: N = 461; Study 2: N = 444) found that, for the same hotel review, third-party reviewers were perceived more favorably than self-reviewers (i.e., hotel managers), leading to more positive evaluations of the hotel. Reviews attributed to AI writers negatively impacted impressions of the reviewer and diminished the review’s impact, with third-party reviewers more negatively affected by AI writer labels than self-reviewers. AI labels of user-generated content can therefore move people out of their default impression formation state. We demonstrate theoretical advancements for warranting theory and AIMC.