Currents of Inquiry: Insights From Two Years of Real-World AI-Learner Water Conversations

Applying a Learning Engineering framework to informal STEM education, this study investigates the alignment between user inquiry patterns and conversational AI responses. We present an analysis of two years of interactions between our Waterbot and community users asking about water issues. Treating questions as signals of learners’ cognitive focus, we code question types and quantify readability, causal connectives, and lexical markers of certainty for both users’ questions and the AI’s responses. Overall, learning sessions remained short, with learning depth tied more to chatbot verbosity than to linguistic complexity or certainty. These findings offer an initial design baseline for conversational AI that supports users’ learning regarding complex issues. The observed asymmetries in length, complexity, and certainty point to concrete levers for the next design cycle, such as briefer default answers and clearer pathways for follow-up questions. As a study grounded in real learner traces, this work establishes an empirical baseline for the next engineering design cycle of conversational systems aimed at civic and scientific informal learning regarding water issues.