| Title |
Designing safe hybrid AI–XR simulations for healthcare communication and interaction training |
| Authors |
Mikkonen, Kristina ; Butkeviciute, Egle ; Armalis, Patrikas ; Aluzaite, Evelina ; Baranauskaite, Rita ; Paulauskas, Lukas ; Spirgiene, Lina ; Subocius, Andrėjus ; Blažauskas, Tomas ; Riklikiene, Olga |
| DOI |
10.1007/978-3-032-28812-7_21 |
| ISBN |
9783032288110 |
| eISBN |
9783032288127 |
| Full Text |
|
| Is Part of |
Digital health and wireless solutions: towards trustworthy and person-centric digital health: 2nd Nordic conference, NCDHWS 2026, Oulu, Finland, June 16–17, 2026: proceedings, Part 1 / M. Särestöniemitric, D. Singh, E. Jarva, J. Reponen (eds).. Cham : Springer, 2026. p. 289-299.. ISBN 9783032288110. eISBN 9783032288127 |
| Keywords [eng] |
Conversational AI ; Extended reality ; Healthcare education ; Hybrid intelligence ; Stress-aware learning |
| Abstract [eng] |
Effective healthcare education must address not only clinical knowledge but also communication, teamwork, and decision-making under pressure. Extended Reality (XR) combined with artificial intelligence (AI) offers new opportunities for immersive, scalable communication training, yet raises critical concerns related to safety, reliability, and learner trust. This paper presents the design rationale and early insights from VirtualHealEd, a human-centred hybrid AI–XR simulation for healthcare communication and interaction training. The platform integrates conversational AI, speech recognition, and immersive XR while employing a hybrid architecture that combines generative AI with deterministic system components to ensure safety in life-critical scenarios. In parallel, multimodal data, including speech transcripts, performance indicators, and physiological signals—are collected to explore stress and cognitive load during AI-mediated interactions with healthcare students (n = 52). Early observations suggest that AI-driven dialogue increases learners’ workload, with substantial inter-individual variability, highlighting the need for adaptive, stress-aware learning designs. The paper contributes design principles for safe hybrid intelligence in healthcare XR education and outlines challenges and future directions for stress-aware adaptive simulations. |
| Published |
Cham : Springer, 2026 |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2026 |
| CC license |
|
