This study examines how Iranians perceive politeness in human–AI interactions, considering cultural norms, social expectations, and technological constraints. Using a mixed-methods approach, thematic analysis of interviews revealed that teachers prioritize status-based politeness, while students value efficiency, reflecting intergenerational shifts. The AI Politeness Questionnaire, developed based on qualitative themes and Beeman’s framework for intercultural politeness, was adapted to assess AI-mediated politeness. The questionnaire, comprising 35 Likert-scale items, was administered online in Persian to university students, teachers, and AI users. Factor analysis revealed four key dimensions of AI politeness: contextual and cultural adaptability, AI’s challenges in processing politeness, user perceptions of AI politeness, and user satisfaction and trust in AI interactions. Findings suggest that AI politeness is a culturally embedded practice, shaped by hierarchical distinctions, indirectness preferences, and generational differences. The study extends Beeman’s analysis of politeness in high-context cultures by integrating AI-mediated politeness and user adaptability, offering deeper insights into politeness in human–AI interactions. These findings emphasize the need for AI systems that can navigate social roles, interpret politeness markers accurately, and align with cultural expectations, ensuring more effective human–AI interactions.