Curious but Unprepared: Healthcare Students’ Perspectives on AI and Robotics in Care and the Need for Curriculum Reform

Abstract

The integration of Artificial Intelligence and Robotics (AI/R) in healthcare presents both opportunities and challenges, especially in developing countries. This study assessed the attitudes and perceptions of Vietnamese healthcare undergraduates towards AI/R applications in healthcare and elderly care. In 2023, a cross-sectional survey was conducted among 1221 Vietnamese healthcare undergraduates. The questionnaire covered demographic, academic, social, and mental factors, as well as attitudes towards AI/R applications measured by a five-level Likert scale. Key findings revealed that respondents were primarily majoring in medicine (60.9%) and pharmacy (29.4%). Awareness and interest in AI/R were high (89.9% and 88.3%, respectively), but formal training was significantly lacking (5.9%). A substantial majority (89.9%) expressed a need for AI/R training. Respondents perceived considerable benefits of AI/R, particularly in data synchronization (mean [M] = 3.83), workload reduction for medical staff (M = 3.79), and delivering multiple healthcare benefits (M = 3.82). Moderate concerns were noted regarding security and privacy (M = 3.46), potential over-reliance on technology (M = 3.43), and AI/R potentially replacing medical staff (M = 3.38). Overall, perceived benefits (M = 3.67) outweighed concerns (M = 3.38), (p < 0.001). Additionally, participants aware of AI/R and those planning to study abroad showed greater interest and training needs in AI/R. Higher GPA and self-esteem were associated with a greater interest in AI/R research. The study highlights a significant gap in formal AI/R training, not only in availability but also in the absence of structured, outcome-based curricula, despite the strong interest among healthcare students in acquiring knowledge and skills in this area. These findings suggest the need for enhanced educational programs to train healthcare students with the necessary competencies to apply AI/R technologies effectively.

1. Introduction

The rapid advancement of Artificial Intelligence (AI) and Robotics (R) is revolutionizing various sectors, including healthcare. Artificial Intelligence/Robotics (AI/R) applications in healthcare promise to enhance diagnostics, administrative tasks, and patient care by providing advanced tools for medical professionals [1,2]. However, the successful integration of these technologies into health systems, particularly in developing countries, poses significant challenges including limited infrastructure, legal liability concerns, data privacy issues, and a lack of adequately trained personnel [3,4,5] and requires a well-prepared workforce, especially in case of medical legal liability [6]. A systematic review about preferences for the use of AI in healthcare has shown that patients, the public, and professionals possessed a positive attitude towards AI healthcare applications [7].

AI/R is at the initial stage of being incorporated into medical education [8,9], but significant progress has only been made in recent years [10]. However, there is limited knowledge about integration of AI/R content into higher education training programs. A disparity has been reported between medical students’ enthusiasm and the training programs provided by the universities [11,12]. Despite the growing body of literature on AI/R in medical education, existing studies are disproportionately focused on medical students in high-income countries, leaving a significant gap in our understanding of how students in other health disciplines, such as nursing, pharmacy, and traditional medicine in low- and middle-income contexts, perceive and engage with these technologies.

Vietnam, like many developing countries, is experiencing a growing interest in incorporating AI/R into its healthcare infrastructure. This interest is driven by the potential benefits of AI/R, such as improving healthcare delivery, reducing the burden on medical staff, and addressing the needs of an aging population. Efforts have been taken in updating curricula with new trends and technologies [13].

Despite these promising prospects, there is limited research on the awareness, attitudes, and preparedness of healthcare students in Vietnam regarding AI/R technologies. Exploring these factors is essential for developing targeted education initiatives that can equip future health professionals with the necessary skills and knowledge to apply AI/R effectively. Taking these evidence gaps, in 2023, a pilot study on a small sample size was conducted on exploring interest and educational gap in AI/R usage for elderly care in healthcare undergraduates [14]. We hypothesized that background demographic, academic, mental, and social factors may contribute to the interest and training needs of AI/R (Figure 1). While the study confirmed a high level of interest and perceived need for training, its small sample size limited generalizability, and it did not detect significant associations between background characteristics and AI/R interest. These limitations highlighted the need for an expanded, more representative investigation to clarify influencing factors and support curriculum development. Based on previous findings and conceptual reasoning, we hypothesize the following: (H1) students with prior awareness of AI/R will report higher interest and training needs in AI/R; (H2) higher academic performance (Grade Point Average, GPA) and self-esteem will be positively associated with interest in AI/R and willingness to engage in AI/R-related research; (H3) students with plans to study abroad will report greater interest and perceived training needs in AI/R.

The objectives of this study are as follows: (1) to explore their awareness and attitudes towards benefits and concerns of AI/R applications in healthcare; and (2) to identify their training needs and interest in AI/R-related research. The current study will provide insights into the current state of AI/R education among healthcare students in Vietnam. Additionally, it will highlight areas where educational interventions are needed to enhance students’ readiness to integrate AI/R technologies into their future professional practice. By addressing these educational gaps, this research aims to contribute to the development of a competent and technologically savvy healthcare workforce capable of meeting the demands of modern healthcare systems.

2. Materials and Methods

A cross-sectional survey was carried out with undergraduate students from a Vietnamese public university specialized in healthcare in 2023. The sample size was calculated using the Cochran formula for categorical data, assuming a 50% proportion for maximum variability, a 95% confidence level, and a 3% margin of error. This yielded a required sample size of 1067 students. Invitations were distributed via institutional email lists and in-class announcements, with access to the online survey (Google Forms) provided through a voluntary sign-up link. The responses were collected over a four-week period. All undergraduate students enrolled in health-related programs (including medicine, pharmacy, nursing, traditional medicine, preventive medicine, and medical technology) were eligible to participate. There were no exclusion criteria beyond voluntary consent. A total of 1500 students were invited to participate in the online survey, and 1221 completed the questionnaire, resulting in a response rate of 81.4%. The questionnaire was developed based on the one used in a previous pilot study [14], which was informed by the published literature on AI/R perceptions in medical education and itself was developed based on published instruments from studies examining AI in medical and healthcare education [6,9,11,12,15,16]. The adaptation process involved the following: (1) a targeted literature review to identify relevant constructs (awareness, interest, perceived benefits, concerns, attitudes towards elderly care applications, training needs, and socio-demographic/academic characteristics); (2) expert review by two health technology and medical education specialists to ensure content validity and alignment with study objectives; and (3) refinement for clarity, cultural appropriateness, and relevance to the Vietnamese healthcare education context. The instrument was reviewed by two education and health technology experts for content relevance and translated into Vietnamese using forward-backward translation. It was administered bilingually (English and Vietnamese) to support comprehension and inclusivity among participants. A pilot test with 30 students confirmed clarity, comprehension, and cultural appropriateness. Internal consistency was high across scales: Perceived Benefits (Cronbach’s α = 0.94), Concerns (α = 0.89), and Self-Esteem (α = 0.84).

The instrument consisted of six sections: (1) demographic, academic, and social characteristics; (2) mental factors measured by a four-point Likert scale based on the Rosenberg scale for assessing self-esteem [17]; (3) awareness, interest, and training experiences regarding AI/R; (4) perceived benefits of AI/R (8 items); (6) concerns about AI/R (5 items); and (6) attitudes toward AI/R applications in elderly care (20 items). All attitudinal items were measured on a five-point Likert scale, where 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, and 5 = Strongly Agree. GPA categories were self-reported based on the institutional grading scale and were interpreted by students as follows: Lower average (≤6.0), Average (6.1–6.9), Upper average (7.0–7.4), Good (7.5–7.9), Upper good (8.0–8.4), and Excellent (≥8.5) on a 10-point scale.

In the survey, AI/R was presented as a broad category encompassing both software-based AI systems and hardware-integrated robotics relevant to healthcare practice. To aid respondent comprehension and ensure consistent interpretation, the questionnaire provided concrete examples of AI applications—such as electronic medical record data synchronization, diagnostic decision support, patient monitoring, and medication reminders—and robotic applications—such as patient lifting, cleaning and sanitization, and automated food delivery. These examples were drawn from the existing literature on healthcare AI/R [3,7,18] and reflected use cases that are feasible or emerging in the Vietnamese healthcare context.

Descriptive statistics were used to analyze participant characteristics. Spearman’s correlation was employed to assess associations between independent variables and training interest in AI/R. A paired samples t-test was conducted to compare perceived benefits and concerns. Cronbach’s Alpha was calculated to evaluate internal consistency. All analyses were performed using IBM Statistical Package for the Social Sciences (SPSS) version 27. Descriptive statistics for Likert-scale items are reported as means (M) and standard deviations (SD), as the data were approximately normally distributed based on skewness and histogram inspection. For variables exhibiting mild skewness, both the mean and median are reported for completeness. Participants were informed about the study’s purpose, and participation was voluntary, confidential, and anonymous.

3. Results

The characteristics of the respondents are shown in

Table 1. Participants demographic, academic, socioeconomical and mental characteristics (N = 1221).

Category	Variable	n	%
Demographic	Sex
	Female	812	66.5
	Male	409	33.5
	Grade
	1st Year	8	0.7
	2nd Year	619	50.7
	3rd Year	378	31.0
	4th Year	95	7.8
	5th Year	97	7.9
	6th Year	24	2.0
Academic	Major
	Medicine	743	60.9
	Traditional Med.	44	3.6
	Nursing	34	2.8
	Public Health	0	0.0
	Preventive Med.	4	0.3
	Pharmacy	359	29.4
	Technician	36	2.9
	GPA
	Lower average (≤6.0)	153	12.5
	Average (6.1–6.9)	375	30.7
	Upper average (7.0–7.4)	344	28.2
	Good (7.5–7.9)	280	22.9
	Upper good (8.0–8.4)	54	4.4
	Excellent (≥8.5)	15	1.2
	English Level (perceived)
	Basic	1007	82.5
	Intermediate	190	15.5
	Advanced	24	2.0
	2nd Foreign Language (Chinese, Japanese, French, Korean, Arabic, Thai…)
	No	1074	88.0
	Yes	147	12.0
Socioeconomical	Marital status
	Unmarried	1211	99.2
	Married	10	0.8
	Lodging status
	Dormitory	225	18.4
	Home	135	11.1
	Rental alone	459	37.6
	Rental shared	402	32.9
	Scholarship
	No	1102	90.3
	Yes	119	9.7
	Part-time Job
	No	822	67.3
	Yes	395	32.4
	Missing	4	0.3
	Family Income (perceived)
	Low	399	32.7
	Middle	816	66.8
	High	6	0.5
	Study Abroad Plan
	Will go	41	3.4
	May go	171	14.0
	May go later	468	38.3
	Will not go	541	44.3
Mental	Self-esteem (Rosenberg)
	M = 26.13	Median = 25.00	SD = 3.102
	Reliability	Cronbach’s Alpha	N of Items
		0.841	10

. Two thirds (66.5%) of the participants are female. The highest group is in their 2nd year (50.7%). More respondents are majoring in medicine (60.9%) and pharmacy (29.4%). The most common GPA category is “Average” (30.7%), followed by “Upper average” (28.2%). Regarding English proficiency, 82.5% perceived to have a basic level, while 12.0% were studying a second foreign language besides English. Most of the population is unmarried (99.2%) and do not receive any scholarship (90.3%). Most live in rental accommodation, either alone (37.6%) or shared (32.9%). A significant portion do not have a part-time job (67.3%). The majority perceive their family income as middle-level (66.8%), with a notable portion perceiving it as low (32.7%). A large proportion do not plan to study abroad (44.3%), while 38.3% may consider it later. The average self-esteem score among the participants is 26.1 (average), with the median score slightly lower at 25.0, suggesting that the distribution is slightly skewed. The standard deviation of 3.10 indicates that the scores are relatively close to the mean. The reliability, measured by Cronbach’s Alpha (0.84), suggests that the scale is reliable in this population.

Table 2. Awareness, interest, and training gaps (N = 1221).

Item	Answer	n	%
Heard about AI/R application in healthcare?	No	165	13.5
	Yes	1056	86.5
Interested in AI/R applications in healthcare?	No	132	10.8
	Yes	1089	89.2
Received training on AI/R in healthcare?	No	1111	91.0
	Yes	110	9.0
Need training about AI/R in healthcare?	No	186	15.2
	Yes	1035	84.8

presents the respondents’ awareness, interest, training, and perceived training gaps related to AI/R in healthcare. A substantial proportion (89.9%) have heard about AI/R in healthcare, showing a high awareness. A large proportion (88.3%) of the participants expresses interest in AI/R in healthcare. Only a small part (5.9%) of the respondents have had training on AI/R in healthcare, indicating a gap in education. Moreover, a major part (89.9%) of the participants expresses a need for receiving AI/R healthcare training, showing a gap that they want to be addressed.

Table 3. Perceived benefits of AI/R (N = 1221).

	Mean	Median	Mode	SD
Synchronizing Data	3.83	4.00	4	0.917
Multiple Benefit	3.82	4.00	4	0.894
Reducing Workload	3.79	4.00	4	0.898
Reducing Waiting Time	3.73	4.00	4	0.898
Support Diagnosis	3.64	4.00	4	0.909
Basic Procedure	3.63	4.00	4	0.925
Reducing the Cost of Treatment	3.44	3.00	3	0.987
Treatment	3.41	3.00	3	0.943

Note: Scale range = 1 (Strongly Disagree) to 5 (Strongly Agree).

illustrates the perceived benefits of AI/R. Overall, respondents view AI/R as highly beneficial, particularly in terms of assisting with data synchronization inside hospitals and across networks (mean [M] = 3.83), which corresponds to a response between “Agree” and “Strongly Agree”. They also recognize that AI/R offer multiple advantages to healthcare (M = 3.82) and can significantly reduce the workload of medical staff (M = 3.79), enabling them to focus on specialized jobs.

Table 4. Concerns of AI/R (N = 1221).

	Mean	Median	Mode	SD
Confidentiality	3.46	3.00	3	0.84
Technical Dependance	3.43	3.00	4	0.89
Replace Staff	3.38	3.00	3	0.93
Reducing Human Contact	3.33	3.00	3	0.96
Inequal Accessibility	3.29	3.00	3	0.92

Note: Scale range = 1 (Strongly Disagree) to 5 (Strongly Agree).

presents the concerns expressed by respondents regarding the use of AI/R in healthcare. The data reveals a moderate level of apprehension of several issues such as privacy and security of patients and themselves (M = 3.46), which falls between “Neutral” and “Agree”, indicating moderate concern, the potential for excessive reliance on technology affecting patient and healthcare staff autonomy (M = 3.43), and the possibility of AI/robotics replacing doctors and medical staff in the future (M = 3.38).

Table 5. Comparison of perceived benefits and concerns (N = 1221).

	Mean	SD	Cronbach’s Alpha	N of Items
Benefit items	3.67	0.78	0.94	8
Concern items	3.38	0.76	0.89	5

Paired samples t-test. t(df) = 12.938(1220); p < 0.001.

compares the perceived benefits versus concerns associated with the AI/R application in healthcare. Although respondents express an average level of concern (M = 3.38), the perceived benefits (M = 3.67) are significantly greater t(1220) = 12.94, p < 0.001, as indicated by a paired samples t-test (p < 0.001). The high internal consistency of the two scales indicates their reliability.

Table 6. Associations of the AI/R needs with various factors (N = 1221).

	Need Training	Interest in AI/R	Like Research AI/R
Heard about AI/R	0.159 **	0.233 **	0.084 **
Trained in AI/R	0.038	0.054	−0.085 **
Sex	−0.032	−0.055	−0.033
Grade	0.030	−0.002	0.033
Major	0.019	0.007	0.000
GPA	0.026	0.004	0.094 **
English proficiency	−0.009	0.039	0.048
Second foreign language	0.017	0.040	0.019
Marital status	−0.012	−0.027	−0.015
Co-living status	0.026	0.041	0.022
Scholarship status	−0.022	−0.090 **	0.003
Part-time job status	0.046	0.033	−0.015
Family income	0.028	0.002	0.036
Study abroad plan	−0.129 **	−0.072 *	−0.122 **
Rosenberg	0.045	0.061 *	0.099 **

Spearman correlation. * p < 0.01; ** p < 0.001.

suggests that participants who have heard about AI/R and who plan to study abroad tend to have interest and training needs and want to do research on AI/R. Individuals with higher GPAs tend to want to pursue research about AI/R. Participants with higher self-esteem also tend to become interested in AI/R and want to conduct research on AI/R. The results show no significant relationship of sex, grade, major, language proficiency, marital and living status, or income and part-time job.

The data on respondents’ attitudes towards elderly care applications of AI/R highlights the most favored functionalities in these technologies (

Table 7. Attitude towards elderly care applications of AI/R (N = 1221).

	Mean	Median	Mode	SD	AI/R
Supports turning on/off lights and surrounding electronic devices	3.76	4	4	0.81	AI
Remind patients according to specific schedules (e.g., taking medicine)	3.75	4	4	0.84	AI
Provide useful information about weather, news, and community services for patients	3.75	4	4	0.80	AI
Support communication with doctors or medical staff	3.75	4	4	0.80	AI
Support lifting heavy objects	3.74	4	4	0.83	R
Clean and sanitize patient room space or patient’s home	3.73	4	4	0.83	R
Monitor security of surrounding areas	3.71	4	4	0.83	AI
Monitor vital signs continuously	3.70	4	4	0.82	AI
Deliver food to patients	3.70	4	4	0.82	R
Monitor food intake (by calories and nutrients)	3.69	4	4	0.80	AI
Support online shopping, ordering, delivery of goods and necessities	3.69	4	4	0.85	AI
Social connection (video call with family)	3.68	4	4	0.82	AI
Assist in moving bedridden patients	3.66	4	4	0.82	R
Escort and assist patients in moving or walking	3.65	4	4	0.80	R
Detect falls and signal assistance	3.62	4	4	0.83	AI
Create a personalized nutritional regimen	3.61	4	4	0.81	AI
Prepare meals according to regimes and recipes	3.60	4	4	0.83	R
Design social activities for patients	3.60	4	4	0.82	AI
Being a companion (telling stories, reading books…)	3.60	4	4	0.85	AI
Provides medical consultation and support services	3.59	4	4	0.81	AI
Check the pain level and call for help when necessary	3.58	4	4	0.81	AI
Designing dementia reduction activities	3.56	4	4	0.81	AI
Organize/participate in recreational activities with patients	3.56	4	4	0.83	AI
Assess mental status (emotions, stress, anxiety…)	3.52	4	3	0.84	AI
Physical care (e.g., personal hygiene, diaper changes)	3.47	4	4	0.84	R

). Among AI applications, the highest mean scores are seen in tasks such as supporting the turning on and off of lights and surrounding electronic devices (M = 3.76), reminding patients according to specific schedules like taking medicine (M = 3.75), providing useful information about weather, news, and community services for patients (M = 3.75), and supporting communication with doctors or medical staff (M = 3.75). These scores indicate a strong preference for AI in enhancing convenience, ensuring timely reminders, providing valuable information, and facilitating communication for the elderly. For robotics applications, the top-rated functionalities include supporting the lifting of heavy objects (M = 3.74), cleaning and sanitizing patient room spaces or homes (M = 3.73), and delivering food to patients (M = 3.70). This suggests a significant reliance on robotics for physical tasks that aid in the daily living and health maintenance of elderly individuals. Overall, the data underscores a positive attitude towards incorporating both AI/R in elderly care, with AI excelling in informational and reminder functions, while robotics is favored for physical assistance.

4. Discussion

4.1. Awareness and Interest in AI/R

The findings of the current study indicate a high level of awareness (89.9%) and interest (88.3%) among respondents regarding AI/R in healthcare. This finding is consistent with numerous studies conducted on students with different majors in various countries [6,11,15,19,20]. This widespread awareness suggests that information about AI/R technologies is effectively reaching the healthcare student population. The substantial interest underscores the perceived potential and relevance of these technologies in the future of healthcare. Compared to counterparts in high-income countries, Vietnamese students appear to report even stronger enthusiasm for AI/R despite receiving less formal training. This contrast highlights a global inequity in AI/R education access, as students in lower-resource settings may be more eager to adopt innovations but lack institutional infrastructure or curricular offerings to do so effectively [6,15].

4.2. Training and Perceived Training Needs

Our results show that despite the high levels of awareness and interest, only a small fraction (5.9%) of respondents have received formal training in AI/R, highlighting a significant gap in the current educational curriculum. The overwhelming majority (89.9%) expressing a need for AI/R training suggests that students recognize the importance of these skills and are eager to bridge this gap. This position aligns with existing evidence regarding the gaps in medical education [12]. These training gaps have also been reported among undergraduate medical students [15,21], nursing students [22], pharmacy students [16], graduate students [11] and professionals [20] from various settings. In all, 66.4% of respondents believed that AI should be taught at the undergraduate level [11,23]. This finding points to a need for educational institutions to revise and integrate AI/R training into their training curricula [9]. This is consistent with the prevailing trend that medical education should evolve to embrace the era of AI [24].

4.3. Perceived Benefits and Concerns of AI/R in Healthcare

Respondents perceive considerable benefits from AI/R technologies, particularly in data synchronization (M = 3.83), enhancing multiple healthcare aspects (M = 3.82), and reducing the workload of medical staff (M = 3.79). These perceptions align with the potential of AI/R to streamline healthcare operations and improve efficiency, allowing medical staff to focus on more complex and critical tasks. This positive perception is well aligned with the existing research literature [7]. Moreover, students believed that AI would benefit their careers [25].

Previous studies have revealed commonly perceived risks included data privacy, reduced work autonomy, algorithmic bias, health inequities, and greater workload to acquire AI skills [7].The current study reveals moderate concerns regarding the security and privacy of patients (M = 3.46), potential over-reliance on technology (M = 3.43), and the fear of AI/R replacing medical staff (M = 3.38). The found low concern about legal liability in case of a medical error is consistent with a previous report [6]. These concerns reflect the need for addressing ethical and practical implications of AI/R implementation to ensure that these technologies augment rather than diminish the role of healthcare professionals. These results align with the fact that healthcare professionals generally view AI positively. The applications of AI in healthcare present promising opportunities but also introduce challenges [26].

It has been reported that students see AI as a partner rather than a competitor [25], and professionals indicated that AI would not replace them in their profession [7]. The comparison of perceived benefits (M = 3.67) and concerns (M = 3.38) reveals that the benefits of AI/R outweigh the concerns (p < 0.001). This positive balance suggests a generally favorable attitude towards AI/R adoption among healthcare students, provided that their concerns are adequately addressed through proper guidelines and safeguards. This result is consistent with a previous pilot study [14]. This optimism may reflect aspirational thinking among students who view technology as a vehicle for innovation and improved patient care. In the Vietnamese context, strong national messaging around digital transformation and the modernization of healthcare may further shape these favorable perceptions. Students likely associate AI/R with professional opportunity more than disruption, particularly in a setting where healthcare resources remain stretched and automation is often framed as a solution.

4.4. Factors Influencing Interest and Training Needs

The data show that participants who have prior awareness of AI/R and those planning to study abroad exhibit higher interest and training needs in AI/R. Higher GPA and self-esteem are also associated with greater interest and desire to engage in AI/R research. These associations highlight the influence of academic and personal factors on students’ attitudes towards AI/R, suggesting targeted strategies for enhancing AI/R education among different student demographics. These findings add new perspectives about the current literature. It has been reported that income, educational background, and experience with technologies were predictors of attitude of pharmacy students toward using AI in pharmacy practice [16]. Among nursing students, factors associated with students’ readiness to apply AI included hi-tech proficiency, knowledge of AI fundamentals, and perceived AI application in nursing practice [22]. It also has been reported that knowledge of healthcare students significantly differed depending on gender, major, experience, and income. The positive attitudes vary by year level and income, and the proportion with favorable perspectives regarding AI varies by age and major [25].

4.5. Attitudes Towards AI/R in Elderly Care

The analysis of attitudes towards AI/R applications in elderly care reveals a strong preference for AI in tasks that enhance convenience and communication, such as turning on/off lights (M = 3.76), reminding patients to take medicine (M = 3.75), and providing information (M = 3.75). Robotics, on the other hand, are favored for physical assistance tasks like lifting heavy objects (M = 3.74), cleaning (M = 3.73), and delivering food (M = 3.70). These preferences indicate the complementary roles of AI/R in supporting elderly care, with AI excelling in informational and reminder functions and robotics in physical tasks. These findings consolidate the high interest in applying robotics for elderly care [18] as well as receiving training on this new discipline [27].

To address the educational gap, we recommend a multi-tiered approach to AI/R training in healthcare curricula. Introductory elective courses on AI fundamentals could be offered in the early years of study, while more advanced online modules and case-based simulations could be incorporated in clinical years. Simulation-based training and interdisciplinary workshops involving computer science and healthcare faculties may also foster practical understanding. These modular approaches offer flexibility and scalability, particularly in resource-limited settings, and align with emerging best practices in digital health education [9,11].

The current study suggests a critical opportunity for Vietnamese health education institutions to act proactively. Integrating AI/R content through elective courses, interdisciplinary workshops, and online learning modules would address the evident student demand. Moreover, embedding AI/R training across both foundational and clinical years would support skill development aligned with future healthcare needs.

This research has limitations. The cross-sectional design captures perceptions at a single point in time, which may not reflect changes over time or the impact of ongoing educational interventions. As the study was conducted at a single university in Vietnam, the findings may not be generalizable to all healthcare student populations in Vietnam or other countries. Institutional culture, curriculum differences, and regional technological exposure may influence students’ perceptions and needs. While multivariable regression would provide deeper insight into independent predictors of AI/R engagement, it was not performed in this exploratory study due to limited outcome variance and unbalanced categorical predictors. The findings should therefore be interpreted as preliminary associations rather than causal relationships. Future studies should aim to include multi-institutional samples for broader representation. Additionally, self-reported data may have biases. The absence of qualitative data limits our ability to explore deeper motivations or contextual factors underlying students’ perceptions. Finally, while the study highlights the need for AI/R training, it does not assess the specific content or effectiveness of existing curricula, leaving a gap in understanding how best to implement AI/R education in healthcare programs.

5. Conclusions

This study highlights the significant interest and awareness among Vietnamese undergraduates regarding AI/R applications in healthcare, despite a notable gap in formal training. The findings reveal a strong perceived need for AI/R education, with students expressing enthusiasm for integrating these technologies into their future practice. The disparity between students’ interest and the current educational offerings underscores the urgent need for universities to enhance their curricula to include comprehensive AI/R training. By addressing this gap, educational institutions can better prepare future healthcare professionals to effectively utilize AI/R technologies, ultimately improving healthcare delivery and patient outcomes. This study provides valuable insights for policymakers and educators aiming to develop targeted strategies for integrating AI/R into healthcare education.

In light of these findings, healthcare curricula in Vietnam and similar contexts could integrate both foundational and application-specific AI/R content. Foundational modules might include AI fundamentals, data analytics, machine learning principles, and ethics of AI in healthcare, while applied modules could address specific technologies such as clinical decision-support systems, medical imaging AI, hospital automation tools, and patient-assistive robotics for elderly care. To ensure effective learning, pedagogical practices should combine didactic lectures with experiential methods, including simulation-based training using AI-enabled platforms, case-based learning with real-world clinical scenarios, and interdisciplinary workshops involving healthcare, computer science, and engineering faculties. Embedding these components into both pre-clinical and clinical stages would help students build not only technical competence but also critical judgment in applying AI/R responsibly in practice. Such curriculum reform can bridge the current training gap and prepare future healthcare professionals to lead the safe and effective adoption of AI/R technologies.