AI disagreement may shake patient trust in doctors
UNIVERSITY PARK, Pa. — Patient trust in medical professionals might hinge on what artificial intelligence (AI) has to say, according to a team led by Penn State researchers.
Using an AI chatbot roleplaying as a human doctor, the team identified how people perceive medical professionals when they think a “human” doctor consults an AI system for a second opinion during mental health consultations. They found that when the AI system agreed with the proxy doctor’s recommendation, patients viewed the professional’s assessment as more credible. But when the AI disagreed with the doctor’s assessment, patient perceptions of medical uncertainty and doctor laziness increased.
The researchers reported their findings in the June/July issue of the International Journal of Human-Computer Studies.
“Historically, when a doctor told a patient that they’re welcome to get a second opinion, they meant ‘go to another doctor’,” said S. Shyam Sundar, Evan Pugh University Professor and the James P. Jimirro Professor of Media Effects at Penn State. “But now in the same session, physicians can use an AI assistant to give a second opinion. We wondered how that might affect patients’ impressions of doctors. We found that perceptions of doctor credibility and medical certainty increase or decrease based on whether the AI assistant agrees or disagrees with the doctor’s diagnosis.”
Recruiting a team of human doctors to give tens or hundreds of patients a consistent experience in terms of clinical practice, communication style and interaction patterns in a controlled experimental setting is impractical, according to the researchers. So, they developed an AI-powered chatbot that could personalize conversations with patients, and gave it instructions to roleplay as a doctor named Dr. Alex.
Then the researchers recruited 135 adults in the United States and offered them an online therapy session with Dr. Alex to identify and discuss daily life stressors. Dr. Alex provided a brief mental health therapy session using the cognitive behavioral therapy (CBT) approach and concluded that CBT is a suitable approach for the patient. Dr. Alex also offered participants the option to get a second opinion from an AI assistant, CareBot. The assistant either agreed or disagreed with Dr. Alex’s recommendation.
After the session, participants answered questions about the doctor’s recommendation, their perceptions and trust in the doctor.
The research team found that agreement between Dr. Alex and the AI chatbot boosted patient confidence in the doctor’s recommendation. Disagreement had the opposite effect, increasing perceptions of medical uncertainty and doctor laziness. Anthropomorphism, or how human-like the AI-powered doctor seemed, also played a role in the observed outcomes. The positive effects of AI agreement and the negative effects of AI disagreement occur only when the doctor was perceived as more human-like.
The findings have important implications for services like telehealth and app-based and other online medical consultations, the researchers said.
“More than half of the participants in our study perceived the AI-simulated doctor as being human-like, demonstrating the capabilities of AI to replicate the professional behaviors of human doctors, at least from a conversational aspect,” said first author Cheng “Chris” Chen, assistant professor of emerging media and technology at Oregon State University who received her doctoral degree from Penn State. “This can bring challenges to doctors who provide online services where their true identity is not visible or clearly communicated. Patients may perceive AI as human or human as AI.”
The effect of AI disagreement on medical uncertainty was especially strong for individuals who believe that machines like AI are more accurate, objective and precise than humans, the researchers noted. These findings suggest that AI can have tremendous authority in planting seeds of doubt in patients’ minds, Sundar said.
The team also suggested strategies for communicating AI disagreement in ways that do not undermine patient trust in their physicians.
“A doctor can still communicate AI disagreement in implicit ways, like saying one of the tools they’re using has highlighted a few points that may be worth exploring a bit more closely, and offering to explore them together,” Chen said.
Explaining potential reasons for AI disagreement can help to reduce medical uncertainty and perceived laziness, Sundar added.
“Calling extra attention to the fact that there’s disagreement and showing that they’re on top of it can reduce the perception of laziness,” he said. “Then if the doctor can lend more nuance into why the medical uncertainty happened — like if the AI system is using data coming from a mostly white, Western sample, and the patient is non-Western, so the results might not apply to them — will communicate the doctor’s activeness.”
Other study co-authors include Yuan Sun, University of Florida, Gainesville; and Mengqi Liao, University of Georgia, who both received their doctoral degrees at Penn State.
Journal
International Journal of Human-Computer Studies
Article Title
When AI Disagrees: The Effect of Second Opinion on Patients' Trust in Doctors
Boston University-led opinion piece highlights the need for public health leadership to build trustworthy AI in health
To realize AI’s promise for health, public health must help shape its future
From disease surveillance to clinical care, artificial intelligence is transforming health. As AI becomes more deeply embedded in decisions affecting population health, a new editorial argues that realizing AI’s full potential requires public health leadership alongside technical innovation.
Dr. Debbie Cheng, assistant dean of data science at Boston University School of Public Health, and lead author of a new peer-reviewed opinion piece published in American Journal of Public Health makes the case that public health leadership is essential to building trustworthy AI in health. Dr. Sandro Galea, dean of the Bursky School of Public Health at Washington University in St. Louis, and Dr. Madhu Mazumdar, director of the Institute for Healthcare Delivery Science at the Mount Sinai Health System in New York City, coauthored the article.
"Public health is a field that has spent decades developing methods to evaluate interventions, monitor impacts, and address inequities," says Cheng, who is also a BU professor of biostatistics and a Hariri Institute Core Faculty member. "Those strengths are highly relevant to AI. It has enormous potential to improve population health, and the greatest opportunity comes when public health expertise helps shape it from the outset."
The editorial asserts that AI should be viewed not only as a technological innovation, but also as a public health opportunity. While AI developers have driven rapid advances, public health brings complementary expertise in understanding how innovations perform in real-world settings, affect different populations, and evolve over time. That perspective broadens how AI should be assessed, looking beyond technical performance to consider trustworthiness, equity, transparency, and population health impact.
The authors highlight examples showing why technical performance alone is not enough. Some AI systems used in dermatology have been shown to perform less accurately on darker skin tones, while other algorithms have underestimated illness severity among lower-income patients because they relied on healthcare costs as a proxy for medical need. Together, these examples illustrate why public health expertise is essential to ensuring AI systems are trustworthy, equitable, and effective across diverse populations.
Beyond evaluating outcomes, Cheng says one of public health's most important contributions is asking critical questions early in the development process: Who benefits? Who might be left behind? What unintended consequences could emerge? Considering these kinds of questions up front, she says, can help ensure AI technologies improve health for everyone.
Cheng is already helping put those principles into practice at Boston University. As founding executive director of the Center for Health Data Science and director of strategy and partnerships for BU's AI Development Accelerator (AIDA), she works with members across the University community to support the responsible integration and adoption of AI in research, education, and administration.
"One of the things I value most about AIDA is the opportunity to bring a public health perspective into conversations about how AI is being adopted across the university," she says. “The best solutions emerge when people from different disciplines learn from one another.”
Looking ahead, Cheng hopes the editorial will encourage greater collaboration among public health researchers, clinicians, educators, and AI developers. One way to achieve that, she says, is by building communities of practice where experts from different disciplines can learn from one another and develop thoughtful approaches to integrating AI into their own work.
More broadly, the authors argue that realizing AI's promise for improving population health will require public health expertise to be embedded throughout the AI lifecycle — from development and implementation to evaluation and long-term governance.
"The future of AI in health will be strongest when technical innovation and public health leadership advance together," Cheng says.
Journal
American Journal of Public Health
Method of Research
Commentary/editorial
Subject of Research
Not applicable
Article Title
Trustworthy Artificial Intelligence in Health Requires Public Health Leadership
Article Publication Date
12-Jul-2026
AI helps nurses stay one step ahead in chronic disease care, new review finds
A comprehensive review finds AI-powered nursing tools improve prediction of health risks and reduce unnecessary hospital use, while more research is needed on patients’ emotional well-being
JMIR Publications
image:
A comprehensive review finds AI-powered nursing tools improve prediction of health risks and reduce unnecessary hospital use, while more research is needed on patients’ emotional well-being.
view moreCredit: JMIR Publications
(Toronto, July 16, 2026) Artificial intelligence (AI) is helping nurses better predict health problems before they become emergencies, according to a new review of existing research published in JMIR Nursing. The study found that AI-based nursing interventions can improve the care of people living with chronic illnesses by identifying patients at greater risk of complications, reducing unplanned hospital visits, and potentially lowering health care costs.
The review, “Effectiveness of Artificial Intelligence–Based Nursing Interventions for Chronic Illness Care: Umbrella Review,” published by JMIR Publications, examined the evidence from eight high-quality systematic reviews.
Lead author Jee Young Joo and colleagues found consistent evidence that AI can support nurses in making more informed, proactive decisions, particularly when caring for people with long-term conditions such as heart disease and diabetes.
Rather than replacing nurses, the findings suggest AI can serve as a clinical decision support tool by analyzing large amounts of patient information to identify warning signs that might otherwise be missed. Across the studies reviewed, machine learning was the most commonly used form of AI.
However, the authors note that there is not yet enough evidence to determine whether AI-based nursing interventions improve patients’ psychological or emotional well-being, highlighting an important gap for future research.
As health care systems face growing numbers of people living with chronic diseases, AI has the potential to become an important part of routine nursing practice. These findings provide guidance for health care leaders and educators looking to integrate AI into nursing care and training.
Original Article:
Joo J, Liu M, Cho Y, Cho H. Effectiveness of Artificial Intelligence–Based Nursing Interventions for Chronic Illness Care: Umbrella Review. JMIR Nursing 2026;9:e97905
URL: https://nursing.jmir.org/2026/1/e97905
DOI: 10.2196/97905
About JMIR Publications
JMIR Publications is a leading independent open access publisher of digital health research and a champion of open science. With a focus on author advocacy and research amplification, JMIR Publications partners with researchers to advance their careers and maximize the impact of their work. As a technology organization with publishing at its core, we provide innovative tools and resources that go beyond traditional publishing, supporting researchers at every step of the dissemination process. Our portfolio features a range of peer-reviewed journals, including the renowned Journal of Medical Internet Research.
To learn more about JMIR Publications, please visit jmirpublications.com or connect with us via X, LinkedIn, YouTube, Facebook, and Instagram.
Head office: 130 Queens Quay East, Unit 1100, Toronto, ON, M5A 0P6 Canada
Media contact: communications@jmir.org
The content of this communication is licensed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, published by JMIR Publications, is properly cited.
Journal
JMIR Nursing
Method of Research
Systematic review
Subject of Research
People
Article Title
Effectiveness of Artificial Intelligence–Based Nursing Interventions for Chronic Illness Care: Umbrella Review
Article Publication Date
15-Jul-2026
Virtual–physical scenario simulation improves nursing students’ learning immersion in home visit training, but skill effects need further study
Virtual-reality simulation boosts immersion, but skill gains remain limited
Home visits are a core part of community nursing. For nursing students, they are not simply a checklist of health assessments. A home visit asks students to observe the family environment, recognize safety risks, communicate with patients and relatives, and turn classroom knowledge into practical care decisions.
Yet such situations are hard to recreate in a classroom. A real home may be cluttered. An older adult may hesitate, question the visitor, become impatient, or respond in unexpected ways. Traditional role-play can help students rehearse the process, but scripted scenarios may not fully capture the uncertainty of entering someone’s home as a nurse.
A study published in the Chinese Journal of Medical Education Research explored whether a virtual–physical integrated scenario simulation could bring students closer to that reality. Researchers from Peking Union Medical College School of Nursing, Beijing Xicheng District Desheng Community Health Center, and Beijing Fengtai District Fangzhuang Community Health Center tested the approach in undergraduate nursing education. They found that the new simulation model increased students’ learning immersion compared with conventional scenario simulation. However, the study did not show a clear short-term advantage in improving students’ home visit attitudes or skills.
The study involved 143 third-year undergraduate nursing students enrolled in 2022 and was conducted from March to April 2025. Students were assigned to either a virtual–physical integrated simulation group or a conventional simulation group. Both groups received the same theoretical teaching on home visits, followed by three hours of experimental teaching. A total of 128 students completed both pre- and post-teaching questionnaires, and 24 students from the virtual–physical group took part in semi-structured interviews.
The teaching task was the same for both groups. Students practiced preparing for a home visit, conducting inquiry and physical assessment, measuring rapid blood glucose, providing health education, and completing post-visit summary, feedback, and documentation. The difference lay in how the scenario was built.
In the conventional group, students worked in a standard laboratory, with classmates role-playing the home visit recipient and family member according to a prepared script. In the virtual–physical group, students trained in a smart eldercare laboratory. The setting combined multimedia equipment, LED displays, physical models, and a wireless smart elderly nursing simulation system. An intelligent simulator played the role of the older adult, while a student played the family member. The system supported AI-assisted inquiry, blood pressure and glucose measurement, and transitions between a community health center and a simulated home environment.
The main quantitative finding was about immersion. Students in the virtual–physical group had a higher total learning immersion score than those in the conventional group: 72.42±8.19 versus 69.44±8.41. The virtual–physical group also scored higher in emotional buy-in and learning experience. These results suggest that the integrated simulation helped students feel more engaged with the learning task.
The interviews help explain why. Many students said the intelligent simulator made the scenario feel more realistic. It could respond to questions, show impatience, ask follow-up questions, or display emotional changes. Some students felt that this made the interaction less predictable than a classmate reading from a script. Others said the simulated home environment, including details such as scattered objects or potential safety hazards, prompted them to observe the surroundings more carefully.
Students also described the experience as more challenging. In conventional role-play, classmates may unintentionally guide or remind each other. In the virtual–physical setting, the simulator had its own response logic. That forced students to listen, adapt, and solve problems more actively. For a teaching activity designed to prepare students for real home visits, this kind of uncertainty may be educationally valuable.
But the same study also cautions against assuming that a more immersive experience automatically produces stronger skills. After teaching, the conventional group showed a small improvement in home visit skill scores compared with before teaching. However, there was no significant difference between the two groups in post-teaching home visit skill scores. The virtual–physical group scored 60.60±6.85, while the conventional group scored 60.89±6.62. The total home visit attitude-skill scores and related dimensions also did not show a clear advantage for the virtual–physical group.
The authors suggest several possible explanations. Home visit skills require repeated practice and feedback. In this study, the experimental teaching lasted only three hours, which may not have been enough time for measurable skill gains to emerge. Students in the virtual–physical group also had to learn how to interact with the smart simulation system. Its novelty and complexity may have drawn some attention away from the core home visit skills being practiced.
Technical limitations also affected the experience. In interviews, students reported delays in system responses, imperfect speech recognition, the need to press buttons during interaction, and occasional breakdowns in conversational flow. Some students felt that conventional role-play, despite being less technologically advanced, allowed more natural and warmer human interaction. The study therefore shows both promise and friction: smart simulation can make training more realistic, but the technology still needs to become smoother and easier to use.
This distinction matters for nursing education. Technology should not be adopted simply because it is new or visually impressive. In home visit teaching, the goal is to help students build observation, communication, assessment, and decision-making abilities. Virtual–physical simulation may support those goals by creating richer and less predictable scenarios, but it requires careful instructional design. Teachers need to keep the learning objectives clear, guide students’ attention to core skills, and provide timely feedback after the simulation.
The study’s findings should also be interpreted with caution. The sample came from a single institution and was relatively small. The home visit attitude-skill questionnaire was designed by the research team and requires further validation. The novelty of the simulation technology may also have influenced students’ learning experience, creating a possible Hawthorne effect. The authors call for larger, multicenter studies using standardized tools and allowing students time to become familiar with the technology before its educational effects are evaluated.
Even with these limits, the study offers a useful lesson for digital transformation in nursing education. Smart simulation can help students feel more present in a home visit scenario. It can expose them to messier communication and environmental details than conventional role-play may provide. But immersion is only the beginning. The next question is how to turn that engagement into durable nursing competence.
Article information：
Article Title: 虚实结合情景模拟在家庭访视教学中的应用效果研究
English Title: Study on the application effects of virtual-reality integrated scenario simulation in home visit teaching
Journal: Chinese Journal of Medical Education Research
DOI: 10.3760/cma.j.cn116021-20251129-02253
Authors: Li Xiaoxue, Zhang Zheng, Wang Lingyun, Wang Li, Guo Aimin
Chinese Journal of Medical Education Research is a monthly peer-reviewed journal sponsored by the Chinese Medical Association and hosted by Chongqing Medical University, under the supervision of the China Association for Science and Technology.
Launched in 2002, the journal publishes research and practice-oriented studies on medical education, with a focus on teaching reform, clinical teaching, curriculum development, residency training, graduate education, nursing education, educational technology, and international medical education. It also features themed columns and special issues on emerging topics and institutional innovations in medical education.
The journal is recognized as a Source Journal for Chinese Scientific and Technical Papers and Citations and is included in the Chinese Science and Technology Core Journals. It is indexed in Wanfang Data, Index Copernicus, the WHO Western Pacific Region Index Medicus, and Ulrich’s Periodicals Directory.
Journal page: http://yxjyts.alljournals.ac.cn/homeNav?lang=zh
Method of Research
Case study
Subject of Research
People
Article Title
Study on the application effects of virtual-reality integrated scenario simulation in home visit teaching
No comments:
Post a Comment