Interesting Points:
- When considering chat interactions with AI, it is important to distinguish between the functional and relational aspects. This study aims to understand the current understanding of the topic.
Summary and Key Insights of the Research: Background and Purpose of the Study: As AI chatbots play a more significant role in everyday life, it becomes crucial to understand how people evaluate communication with AI. This study compares two theoretical perspectives that explain human-AI communication: a) Communication Script Theory (especially the CASA theory) b) Expectancy Violations Theory
Theoretical Framework: Communication Script Theory: People apply human communication scripts to interactions with AI. Expectancy Violations Theory: Communication evaluation is based on whether prior expectations were met or violated.
Research Design and Methods: Pilot Study: Confirmed differences in expectations between humans and chatbots. Main Experiment: Experiment design with 2 (contextual fit) x 2 (response speed) x 2 (communicator’s ID). Dependent variables: Reliability (ability, trustworthiness, benevolence) and attractiveness (social attractiveness, task attractiveness).
Key Findings: Functional aspects (reliability, task attractiveness): Both humans and chatbots were most highly rated when showing high contextual fit and fast responses. Relational aspect (social attractiveness): Humans were consistently rated higher than chatbots. Predictions of Expectancy Violations Theory were not supported.
Theoretical Contributions: Suggests that the Communication Script Theory has greater explanatory power in human-AI communication. Clarifies the existence of functional and relational aspects in evaluating communication with AI.
Practical Implications: Chatbot developers should prioritize high contextual fit and appropriate response speed. However, they should recognize the limitations of AI in social relationship building.
Novelty and Contributions: Comparative examination of theoretical frameworks for evaluating human-AI communication. Clear distinction between functional and relational aspects, highlighting their importance in evaluating AI. Suggests the difficulty of applying Expectancy Violations Theory to human-AI communication.
Future Research Directions: Changes in evaluation during long-term interactions with AI. Further exploration of the interaction between functional and relational aspects. Comparison of human-AI communication evaluations in different contexts and cultures.
This study sheds light on the complexity of human-AI communication and provides important insights for future AI development and understanding of human-AI relationships.