ChatGPT Voice Trains Future Doctors

A new study explores using ChatGPTs advanced voice mode to help medical students practice communication skills. Findings suggest AI can boost confidence and supplement traditional training methods effectively.

Effective communication is a cornerstone of good medical practice and patient care. However, providing medical students with enough opportunities to practice these vital skills can be a real challenge in medical education. Traditional methods, like using standardized patients (SPs), are valuable but come with their own set of hurdles, including high costs and logistical complexities. This has led researchers to explore new avenues, and one exciting prospect is the use of conversational artificial intelligence (AI). A recent study delved into how AI, specifically ChatGPT, could help supplement communication skills training for medical students.

The Challenge of Teaching Medical Communication

Mastering communication is not just about knowing what to say, but how to say it, especially in sensitive or high-stress medical situations. Medical schools often find it difficult to:

• Provide frequent practice sessions for every student.
• Cover a wide variety of patient scenarios.
• Offer immediate, constructive feedback.
• Overcome the financial and logistical burdens of methods like SPs.

This is where technology, particularly AI, might offer a helping hand.

How AI is Changing the Game: The Study

Researchers conducted a mixed-methods study involving 27 medical students from three universities in the UK. The study ran from November 2024 to March 2025. Participants used the freely available ChatGPT, in its advanced voice mode, to act as a standardized patient for role-playing communication scenarios.

The study focused on three key communication domains:

1. Dealing with challenging patients
2. Breaking bad news
3. Counselling anxious patients

Before and after their AI role-play sessions, students completed assessments to measure the perceived usefulness of the AI and their self-reported confidence in these communication skills. The researchers adapted the Immersive Technology Evaluation Measure (ITEM) for this purpose. Quantitative data, like changes in confidence scores, were analyzed using statistical tests, while qualitative feedback from students underwent a detailed thematic analysis.

Key Findings: A Boost in Confidence

The results were quite promising, showing a significant jump in students' self-reported confidence after practicing with ChatGPT:

• Dealing with challenging patients: Median confidence score increased from 3 to 4 (P < 0.001).
• Breaking bad news: Median confidence score rose from 2 to 4 (P < 0.001).
• Counselling anxious patients: Median confidence score improved from 3 to 4 (P < 0.001).

These improvements suggest that AI simulations can be effective in making students feel more prepared for difficult conversations.

Interestingly, the study also found that the university a student attended influenced their willingness to use the AI simulation again (odds ratio (OR) = 0.19) and how realistic they found the AI-generated scenarios (OR = 0.18). This indicates that individual or institutional experiences might shape perceptions of AI's utility.

From the qualitative feedback, four main themes emerged:

• Varied Experiences and Expectations: Students came in with different levels of comfort and anticipation regarding AI.
• Valuable Features: Participants particularly liked the detailed feedback AI could provide and the "safe practice environment" where they could make mistakes without real-world repercussions.
• Limitations Noted: Some technical challenges were reported, along with a sense that the AI had a somewhat restricted emotional range compared to a human.
• AI's Complementary Role: Most students viewed AI not as a replacement for human-based training but as a valuable addition to their educational toolkit.

The Bigger Picture: AI as a Learning Partner

This exploratory study offers early evidence that conversational AI like ChatGPT can enhance medical students' confidence in challenging communication skills. It also highlights AI's ability to provide structured feedback in an accessible way.

However, the researchers advise interpreting these findings with caution. The study had a small sample size, lacked a control group (to compare against traditional methods or no intervention), and relied on students' subjective self-assessments of confidence. Identified limitations, such as the AI's emotional authenticity and occasional transcription inaccuracies, need to be addressed for future applications. There were also notable differences in how students from different institutions perceived the realism of the AI scenarios.

Despite these constraints, the students highly valued the psychological safety and sheer convenience that AI simulation offered. The findings strongly suggest that AI simulation has significant potential as a complementary tool to broaden the opportunities for communication skills practice.

Of course, more research is needed. Larger, controlled studies are essential to firmly establish how effective AI is compared to traditional approaches. Future investigations should also incorporate objective measures of skill improvement (not just self-reported confidence) and explore whether skills honed with AI successfully transfer to real-life clinical interactions with patients.

This study opens a fascinating window into how AI can support the development of crucial soft skills in the medical field, potentially transforming aspects of medical education for the better.

Explore More from the Source

The original research was published on Cureus. If you are interested in learning more about the platform:

• About Channels: Unlock discounted publishing that highlights your organization and the peer-reviewed research and clinical experiences it produces. Learn more
• Academic Channels Guide: Find out how channels are organized and operated, including details on the roles and responsibilities of channel editors. Learn more