How a chatbot should communicate advice? Formal language matters.
(A2024-119898)

AUTHORS

Novin Hashemi, University of Bologna; Gabriele Pizzi, University of Bologna; Chiara Orsingher, University of Bologna; Virginia Vannucci, University of Bologna

ABSTRACT

Abstract: This study explores how chatbots should communicate advice, focusing on language formality and its impact on perceived politeness, reactance, and advice adherence. We propose a conceptual model aligning with politeness theory, Computers As Social Actors (CASA), and the reactance behavior. Our findings suggest formal language significantly enhances negative politeness perception, while informal language enhances positive politeness perception but the P value is .056 slightly higher than 5%. Furthermore, perceived positive politeness increases reactance while perceived negative politeness decreases the effect. Finally, lower reactance contributes to higher advice adherence intention. We conclude that from our study in advice advice-giving people react similarly to human-to-human interaction and politeness can affect their perception however it seems that formal language is more persuasive than informal in this case.