Human-to-AI Interaction in Faith Contexts: Analyzing Conversations with 'Ask Cathy'

​As educators and community leaders increasingly utilize generative AI to create interactive educational activities, understanding the nuances of human-AI conversations becomes essential. Religious dialogues, rich with existential and factual inquiries, provide a unique lens for examining these interactions. This study analyzes over 85,000 messages from more than 10,000 conversations with "Ask Cathy," a chatbot designed to answer questions about the Episcopal Church using Retrieval-Augmented Generation (RAG) techniques. Our research team, comprising experts in linguistics and computer science, employed various methods—including topic and sentiment analysis validated against human-assessor baselines—to assess user engagement and evidence of learning. Preliminary findings reveal distinct patterns in human-AI religious dialogues, offering insights into designing effective chatbots for education and faith formation. These results hold significant implications for educators, technologists, and faith leaders seeking to foster meaningful interactions through AI-driven platforms.