AI Virtual Patients in Audiology Education: Student and Educator Perspectives

INTRODUCTION Clinical competence in audiology requires more than technical knowledge—it demands refined communication skills, empathetic counseling abilities, and the confidence to navigate diverse patient interactions. Patient-centered care must be central to audiology education programs to ensure future audiologists deliver holistic, individualized, and culturally responsive care.1 Research has shown that audiology students would appreciate greater opportunities to refine their clinical communication skills.2 Simulated learning environments (SLEs) or simulation clinics have served as safe learning environments where students practice these essential skills before encountering real patients. To date, SLEs have included the use of role playing, standardised patients, computer-based simulations, and mannequins for learning and assessment in diagnostic and rehabilitative audiology.3 However, SLEs may experience limitations such as scheduling constraints, availability of standardised patients, and variability in case exposure.The use of Artificial Intelligence (AI) and the creation of virtual patients offers an opportunity to address some of these limitations. In 2025, The University of Queensland (UQ) Masters of Audiology Studies program in Australia collaborated with the National Acoustics Laboratories to adapt the NAL-VP tool for educational purposes. NAL-VP is an AI-powered simulation tool that enables students to conduct realistic consultations with virtual clients through voice- or text-based interactions, using OpenAI’s GPT-4o large language model. The platform is web-based and runs through standard modern browsers (e.g., Chrome, Safari, Edge) on PCs, laptops, and mobile devices. It requires an active internet connection and no local installation. NAL-VP generates dynamic, interactive responses based on carefully designed case profiles; no identifiable patient data are used. UQ provided 10 patient cases from the adult simulation clinic, each with varied case histories and hearing conditions. The resulting UQ NAL-VP personas were iteratively refined following pilot testing and feedback from a wide range of users, including audiology students, practicing clinicians, clinical educators, trainers, and hearing industry professionals to ensure authentic hearing histories, personalities, and communication styles. Patient characteristics such as their willingness to provide personal information and openness to accepting recommendations can be modified. Lastly, each session can be transcribed for feedback, reflection, and for assessment purposes. The integration of such AI-powered tools into pre-professional audiology education raises important questions. How do students and educators perceive these tools? Can AI genuinely complement traditional learning methods? What safeguards ensure appropriate use? During the latter half of 2025, we discussed these questions through interviews and short surveys with clinical educators, course coordinators, and final-year audiology students. A number of common themes emerged and are presented here. AI AS ENHANCEMENT, NOT REPLACEMENT The most consistent finding across all stakeholders was clear: AI virtual patients should complement, not replace, traditional simulation clinics and real patient interactions. Students and educators alike viewed generative AI virtual patients as a supportive resource that fills specific gaps in clinical education. The dominant gap identified was in allowing self-directed practice opportunities. Students noted the value in engaging with virtual patients outside scheduled clinic hours, and in practicing case history taking, counseling techniques, and communication flow without the pressure of evaluation or the constraints of limited clinic time. Students who utilized AI practice sessions reported feeling notably more confident and better prepared for both simulation clinics and real-world clinical placements. This supplementary role was identified as particularly valuable for students facing specific challenges: those with limited placement or peer practice exposure, students managing social anxiety in clinical settings, or students from culturally and linguistically diverse backgrounds. The UQ NAL-VP platform was seen to provide a low-stakes environment where practice and learning are possible without time constraints, consuming educator time or requiring coordination with standardised patients or other resources. SETTING THE CORRECT CONTEXT Our participants emphasized that simply providing access to technology without context invites confusion and misuse. It was suggested that effective implementation should incorporate introductory workshops explaining the virtual patient tool’s purpose, capabilities, and inherent limitations. Students need demonstrations showing how to use virtual patients within the simulation clinic framework as one element of comprehensive clinical education. Ongoing support through discussion sessions with clinical educators, reference guides, or instructional videos would help build AI literacy throughout the academic term. Equally important is educator preparation. Clinical supervisors must feel confident in the technology to model appropriate use and supervise student engagement effectively. THE IMPORTANCE OF CLINICAL EDUCATOR FEEDBACK While the UQ NAL-VP tool was seen as providing valuable independent practice opportunities, the guidance of clinical educators in building communication skills remains key. Educators should understand both the potential and the boundaries of AI virtual patients to ensure they can better guide students toward productive learning experiences. Our research revealed potential for misuse when students lack explicit guidance on how to engage with these tools ethically and effectively. Potential concerns named included assuming the virtual patient responses are always correct or authoritative, and copying AI-generated phrases without reflection or genuine understanding. These approaches could undermine the learning objectives the technology aims to support. Appropriate AI use requires active engagement and critical thinking. Students should reflect on their virtual patient interactions, compare virtual patient-generated feedback with educator guidance, and integrate insights across AI sessions and real-world clinic experiences. The transcript function was seen as valuable as a tool for such guidance and reflection. Clinical educators could use virtual patient appointment transcripts in identifying communication patterns, recognizing missed opportunities, and tracking skill development over time. MONITORING, FEEDBACK, AND CONTINUOUS IMPROVEMENT The digital nature of AI platforms enables tracking and analysis impossible with traditional standardized patients. Usage logs can capture student engagement time, session frequency, and common areas of difficulty such as counselling skills, appointment dynamics, or question sequencing. This data serves multiple purposes. Clinical educators can tailor feedback based on individual usage patterns and identify students who may need additional support. Students benefit from reviewing transcripts during self-reflection and educator debriefing discussions. However, our research also highlighted an important challenge: the lack of access to feedback during independent UQ NAL-VP sessions. Student participants wanted feedback on their skills during or at the end of the session. Future NAL-VP tool research and development should investigate the creation of appropriate criteria for standardized feedback. Participants identified other desirable features for future AI development, including detection of students’ vocal tone and loudness, conversational flow analysis, detailed time logs for tracking engagement, flexible personas reflecting diverse client backgrounds, and emotionally responsive virtual patients that adapt to student communication cues. The challenge in developing these features are significant. Creating culturally sensitive and appropriate AI personas requires careful consideration and extensive testing. Defining “standard” communication flow in rehabilitation audiology lacks clear benchmarks and underestimates the importance of the students to practice person-centred care. Perhaps most importantly, students and educators must accept the limitations in the use of virtual patients—the goal for such tools is filling practice and learning gaps, not replicating human spontaneity or replacing genuine clinical practice experiences. TRUST, TRANSPARENCY, AND REALISTIC EXPECTATIONS Trust emerged as a pivotal factor in AI adoption. Educators need confidence that AI produces accurate, contextually appropriate responses aligned with evidence-based practice. Students must believe the tool genuinely helps prepare them for real-world audiology interactions. Participants suggested that building this trust requires transparency about AI capabilities and clear communication about the role of the tool within the learning environment. Both students and educators felt the UQ NAL-VP tool could support a scaffolded learning approach, allowing students to build foundational communication skills in a low-risk environment before progressing to more complex simulations and real patient interactions. EVIDENCE GAPS AND RESEARCH NEEDS All participants highlighted that AI implementation in audiology education currently lacks robust evidence demonstrating its effectiveness. While student and educator perceptions appear positive, research must evaluate whether AI virtual patients genuinely improve clinical competence, communication skills, and reflective practice abilities. Future investigations should explore several critical questions. How do students’ perceptions and understanding of AI tools such as the UQ NAL-VP tool evolve with training and experience? What feedback criteria produce meaningful learning outcomes? How can we establish culturally appropriate and diverse virtual personas that serve all students effectively? Can we reliably assess improvements in vocal tone, counselling approach, or conversation management through AI interaction? Until this evidence base develops, institutions should position AI virtual patients as promising but still-evolving educational tools requiring ongoing evaluation and refinement. PRACTICAL AND PEDAGOGICAL CONSIDERATIONS Integrating AI virtual patients requires careful attention to resources, policies, and ethical frameworks. Financial considerations include licensing fees, platform maintenance, staff training costs, and ongoing technical support. Time investments span initial setup, educator preparation, student orientation, and continuous quality monitoring. When introducing AI-powered tools, institutions should develop clear guidelines addressing ethical use, data privacy, student consent, appropriate integration into curricula, and assessment policies. The use of virtual patients and AI-generated feedback could support formative learning, but its role in summative assessment is also still emerging. CONCLUSION The use of AI virtual patients such as those in the UQ NAL-VP tool represents a valuable addition to the audiology education toolkit when implemented thoughtfully. The tool offers flexible, self-directed practice opportunities that help students build confidence and refine communication skills between traditional clinic sessions. However, success depends on proper framing, adequate training, realistic expectations, and ongoing research into effectiveness. The technology works best as a bridge between classroom theory and clinical practice—a tool that enhances rather than replaces human interaction in audiology education. Maintaining this balanced perspective while building the evidence base will determine the ongoing value of AI virtual patients and their use in audiology education.