Triadic Relations in Healthcare: Surveying Physicians' Perspectives on Generative AI Integration and its role on Empathy, the Placebo Effect and Patient Care

Vanda Faria^1,2*Nathan Goturi²Amanda Dynak²Cameron Talbert²Nicholas Pondelis²Marco Annoni³Charlotte R Blease^4,5Scott A Holmes⁶Eric A Moulton^2,7

• ¹Deaprtment of Psychology, Uppsala University, Uppsala, Sweden
• ²Brain and Eye Pain Imaging Lab, Department of Anesthesia, Critical Care and Pain Medicine, Boston Children's Hospital, Harvard Medical School, Massachusetts, USA, Boston, United States
• ³Bioethics Research Unit - CNR Interdepartmental Center for Research Ethics and Integrity (CID-Ethics) National Research Council, Italy, Milan, Italy
• ⁴Department of Women's and Children's Health, Faculty of Medicine, Uppsala University, Uppsala, Uppsala, Sweden
• ⁵Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, United States
• ⁶Pediatric Pain Pathways Lab, Pain and Affective Neuroscience Center, Department of Anesthesiology, Critical Care and Pain Medicine, Boston Children’s Hospital and Harvard Medical School., Boston, United States
• ⁷Department of Ophthalmology, Boston Children’s Hospital and Harvard Medical School, Boston, United States

Background: The integration of generative artificial intelligence (genAI) tools into clinical practice and health care systems is reshaping modern healthcare, introducing technology as an active third participant in the evolving physician–patient–technology relationship. As these tools begin to play more prominent roles, understanding physicians' perspectives is essential for guiding their ethical and effective use. Objective: This survey examined physicians' use of genAI and their views on its potential impact on empathy, the physician–patient relationship, and psychobiological mechanisms such as the placebo and nocebo effects. Methods: A cross-sectional survey was distributed to 2,444 physicians at a major academic pediatric hospital in Boston (October 2024–February 2025). The survey included items on genAI use, perceptions of its clinical and relational impact, and associated concerns. A total of 319 (13%) completed responses were analyzed using descriptive and inferential statistics. Results: Within two years of the public release, 65.2% of respondents had used genAI tools, most commonly for administrative tasks like writing emails (55.8%) and documentation (31.3%). Use was more common among younger physicians and men. A majority believed genAI could improve patient care (60.7%) and increase time for direct interaction (65.7%). However, views were more divided regarding its ability to support empathetic care or influence placebo effects, with many physicians expressing neutrality or uncertainty. Notably, 50.8% agreed that genAI-human interactions could increase patient anxiety, indicating concern about potential nocebo effects. Perspectives on broader genAI adoption were mixed, with 30.6% expressing concern and 37.7% neutral. Conclusions: Physicians are rapidly adopting genAI tools, primarily for administrative use, while remaining cautious about its relational and psychological implications. These findings underscore the importance of addressing ethical concerns and supporting clinicians as they navigate this evolving triadic relationship between physician, patient, and genAI.