Do You Say “Please” to ChatGPT? Investigation of User Perceptions of AI Consciousness and Their Politeness Performance in Human-Chatbot Conversations

The rapid growth of conversational AI transforms human–AI interactions. Understanding users’ perceptions and attitudes toward AI is crucial but remains underexplored, especially regarding perceived AI consciousness. This research examines the impact of interaction modality (text vs. voice) on users’ perceived AI consciousness and their moral performance toward AI chatbots. To this end, a series of controlled experiments were conducted with 25 participants interacting with ChatGPT in both text and voice modes. Results show that voice interactions significantly increase perceived AI consciousness and lead to greater politeness in users’ performance during conversations with AI chatbots. However, perceived AI consciousness does not significantly affect politeness performance. The findings underscore the role of interaction modality in shaping users’ perceptions and behaviors, highlighting a more direct link between modality and politeness performance. This work provides insights for the design of more socially engaging multimodal AI systems.