Examining user-AI interaction patterns in health-information queries

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: Not reported — no quantitative effect sizes, frequencies, or proportions are provided in the abstract

• Primary Outcome: Thematic categorisation of health-related user-AI interactions into four domains: symptom exploration, medical education, mental health support, and general health advice

• Nnt Or Sensitivity: SVM classifier performance metrics (sensitivity, specificity, F1-score, AUC) not reported in the abstract; cannot be assessed without full-text review

• Confidence Interval: Not reported

Clinical Application

The study does not propose or evaluate a clinical intervention; it is descriptive. No clinical workflow change, tool, or protocol is recommended. Feasibility of applying findings is therefore not directly assessable. The findings have limited direct applicability to Australian clinical practice at this stage. The TGA has published guidance on Software as a Medical Device (SaMD) and AI/ML-based tools, but the GAI chatbots studied here are not identified and may not be TGA-regulated. The RACGP has acknowledged the growing role of digital health tools in patient information-seeking but has not yet issued specific guidance on GAI chatbot use by patients. The PBS is not relevant to this study. Australian clinicians should be aware that patients may be using unregulated AI chatbots for health queries — particularly for mental health support — and should proactively discuss this in consultations. The Australian Digital Health Agency's national digital health strategy is relevant context for interpreting the broader implications of this research. Potentially relevant to clinicians managing patients who use AI chatbots as a first point of contact for health queries, particularly in primary care, mental health, and health literacy contexts. However, the study population is insufficiently characterised to define a specific applicable clinical population.

Abstract

In this study, we examine how individuals utilize generative artificial intelligence (GAI) when seeking health-related information. Using a dataset of user-GAI chat logs available on Hugging Face, we analyzed real-world interactions in which users posed health-related questions to a generative model. We applied a combination of data and text-analytic methods to categorize these interactions, including supervised machine learning techniques such as Support Vector Machines (SVMs). SVMs were selected for their efficiency and strong performance in high-dimensional text classification tasks, and used to identify recurrent themes in user queries and interactions. We found that users frequently consult AI chatbots for symptom exploration, medical education, mental health support, and general health advice. The findings suggest that GAI tools may not only function as informational resources, but also as preliminary support tools that can shape users' health knowledge and encourage them to seek consultation with medical professionals.

References

1. 1.Wang, J. T., Chung, H. W., Do, G. N., & Yang, A. T. (2026). Examining user-AI interaction patterns in health-information queries. International Journal of Medical Informatics. https://doi.org/10.1016/j.ijmedinf.2026.106453