Exploring the Potential of AI Chatbots in Health Behavior Change and Motivational Interviewing

Changing health-related behaviors such as quitting smoking, increasing physical activity, and adhering to treatment plans is often challenging but vital for managing chronic illnesses. Motivational interviewing (MI), a patient-centered counseling approach, has been shown to be effective in fostering motivation and supporting behavior change across diverse healthcare environments.

Despite its proven benefits, MI is underutilized in clinical practice primarily due to obstacles like time constraints, the need for specialized training, and reimbursement issues. Recently, advances in artificial intelligence (AI) are opening new possibilities for delivering MI through digital platforms. AI-powered chatbots, virtual agents, and mobile applications can mimic empathetic and supportive conversations characteristic of MI, providing users with accessible, around-the-clock support that is free of judgment. This approach can be particularly advantageous for individuals who are less likely to seek traditional behavioral health services.

Initial studies indicate that these digital tools are both feasible and acceptable to users. However, questions remain about how well they adhere to MI principles such as empathy and autonomy support, and whether they can produce meaningful, long-lasting behavioral changes. Measuring fidelity to MI—ensuring these AI systems genuinely reflect core counseling techniques—is complex, as current evaluation methods often require detailed human review that does not scale well.

To address these gaps, researchers from Florida Atlantic University’s Charles E. Schmidt College of Medicine conducted a comprehensive review of existing studies on AI-driven systems that aim to deliver MI. Their focus was on understanding how these tools are employed, their usability, adherence to MI principles, and their effectiveness in influencing psychological and behavioral outcomes.

Their review, published in the Journal of Medical Internet Research, found that chatbots were the most common AI tools used in this domain, with some virtual agents and mobile apps also involved. These systems utilize a range of technologies—from rule-based scripts to advanced large language models like GPT-3.5 and GPT-4. While they aim to simulate motivational interviewing, the quality of evaluations varies; most studies did not thoroughly assess safety measures related to AI misinformation or inappropriate responses.

Most research to date has focused on psychological factors such as readiness to change and feeling understood, with few studies examining actual long-term behavioral outcomes. Notably, evidence on sustained behavior change remains limited, partly due to short or absent follow-up periods. Nonetheless, AI tools show promise in delivering motivational content effectively and engaging users early in their behavior change journeys.

Experts underscore that many digital interventions incorporate motivational elements but do not consistently follow structured MI practices or demonstrate rigorous fidelity. User feedback highlights appreciation for the convenience and structure provided by AI systems, though many users miss the emotional depth and relational nuances of human counseling.

The participant groups studied ranged from general adults to college students and patients with specific health conditions, with smoking cessation being the most common focus, followed by substance use, stress management, and other health behaviors.

The authors emphasize that AI-driven tools hold exciting potential to broaden access to motivational interviewing and promote impactful health behavior changes. Future research should focus on robust evaluation of these systems’ effectiveness, transparency, and ethical considerations. By integrating scalable AI technology with empirically supported behavioral frameworks, healthcare providers can better support individuals facing behavior modification challenges.

Study co-authors include FAU medical students Zev Karve, Jacob Caley, Christopher Machado, and senior medical librarian Michelle K. Knecht.

source: https://medicalxpress.com/news/2025-10-chatbots-doc-explore-ai-health.html