Artificial intelligence in nursing education: knowledge, perceptions, and associated factors among Egyptian nursing students across universities

Abstract

Aim

To assess Egyptian nursing students’ level of AI knowledge, perceptions of AI benefits, and fears regarding AI in nursing education and practice, and to examine associated sociodemographic factors.

Background

The rapid integration of artificial intelligence (AI) into healthcare has generated both enthusiasm and apprehension among nursing professionals and students. While AI offers substantial potential benefits in clinical efficiency, decision support, and educational innovation, important concerns remain regarding individualised care, professional displacement, and data privacy. This combination of promise and concern suggests a cautiously receptive context in which nursing students may recognise AI’s value while remaining uncertain about its implications for practice and education. Therefore, evidence-based assessment of nursing students’ knowledge and perceptions is needed.

Methods

A cross-sectional descriptive design was utilised in this study. A convenience sample of 2412 nursing students was drawn from two nursing faculties between May 2024 and January 2025. Data were collected using a structured three-part questionnaire covering individual sociodemographic characteristics, a dichotomous AI knowledge test, and an attitude scale measuring perceived benefits and fears regarding AI.

Results

Approximately 46.4% of students demonstrated good AI knowledge (mean score 3.2 ± 1.3 out of 5). The mean perceived benefits score was 10.9 ± 3.5 (maximum 14), indicating generally favourable views of AI’s educational and clinical support applications. The mean fear score was 5.0 ± 1.6. Notably, 72.2% of students expressed concern that AI may replace nurses in the future, and 57.8% reported discomfort with using AI in educational settings. Statistically significant differences were observed across academic years for knowledge, perceived benefits, and fears, and across age groups for perceived benefits and fears; age-related differences in knowledge were smaller but remained statistically significant. No significant differences were found by gender or university.

Conclusion

Egyptian nursing students demonstrated a foundational understanding of AI and broadly recognised its benefits for education and clinical support. Nevertheless, substantial gaps in formal AI training and pronounced fears about professional displacement were identified. Structured, ethically grounded AI educational programs are urgently needed to enhance students’ digital competencies and ensure the safe and effective use of AI in nursing practice.

Clinical trial registration

Not applicable.

What is already known regarding this topic?

What does this paper add?

The implications of this paper

• Artificial intelligence (AI) is increasingly being integrated into healthcare systems and nursing-related contexts, with potential to improve efficiency, decision support, and selected aspects of patient care.

• At the same time, AI adoption raises important ethical and professional concerns, including data privacy, accountability, and the possible impact of automation on healthcare roles.

• Because nurses play a central role in patient care delivery, their readiness and acceptance are important to the safe and effective integration of AI into practice.

• This study provides evidence from Egypt on nursing students’ knowledge of AI and their perceptions of its benefits and risks in nursing education and practice.

• The findings show that students generally recognised several potential benefits of AI, while also expressing substantial concerns, particularly regarding nurse replacement, data confidentiality, and the appropriateness of AI in educational settings.

• Significant differences in AI-related knowledge, perceived benefits, and fears were observed across academic years, while age was associated with perceived benefits and fears and showed a smaller association with knowledge. No significant differences were found by gender or university.

• Nursing education programs should integrate structured AI-related content into undergraduate curricula to strengthen students’ digital competence and preparedness for AI-enabled practice.

AI education should include ethical and professional dimensions, particularly data privacy, accountability, bias, and the continuing role of human clinical judgment in patient-centred care.

• Educational strategies such as case-based learning, supervised discussion, and critical appraisal of AI-supported outputs may help reduce uncertainty and support informed, responsible use of AI in nursing.

Introduction

Modern healthcare is becoming increasingly reliant on artificial intelligence (AI), a component of the “fourth” digital industrial revolution [1]. Artificial intelligence is increasingly being integrated into healthcare systems to support clinical decision-making, data analysis, and patient care processes, owing to its ability to analyse and interpret vast volumes of data [2]. AI is advancing rapidly, and its use in the medical field is growing. It is gaining popularity across a variety of medical disciplines, including dermatology, pathology, and ophthalmology. Assistive AI tools, such as screen readers and Alexa, are becoming more popular among students and are designed to enhance and support human tasks. Similarly, generative AI can also be utilised for assistive purposes, such as aiding language learning [3]. In this study, artificial intelligence refers to computer-based technologies capable of performing tasks that typically require human intelligence, such as data analysis, pattern recognition, and clinical decision-support applications in healthcare.

As AI applications advance in clinical settings, improving care quality and optimising workflows, nursing education must adapt to equip students to meet the demands of the rapidly evolving healthcare landscape [4]. One branch of artificial intelligence (AI) is machine learning (ML), which involves training machines to identify medical conditions by analysing images, audio, and other data. Machine learning is expected to be indispensable in the future of medicine, given its capacity to analyse the vast quantities and diverse formats of data generated daily by healthcare institutions and patients, thereby contributing meaningfully to clinical decision-making [5].

The healthcare landscape is undergoing a fundamental transformation. Rapidly evolving technologies, shifting demographics, and changing consumer expectations are collectively reshaping how care is delivered globally. Increasing numbers of inpatient facilities are being converted to home care and outpatient services, while critically ill patients continue to require acute inpatient care [6]. Furthermore, AI and ML will be integral to future nurses’ engagement with patients across diverse healthcare settings, unlike in the current landscape [7].

One significant advantage of AI in clinical settings is its ability to enhance nursing performance and patient safety. AI-powered systems can identify patients at the highest risk and deliver pertinent decision-support information to those requiring urgent intervention [8]. Such tools support proactive, precision-based care. However, the integration of AI also poses inherent risks, including concerns about supplanting the human dimension of care, depersonalising interactions, and reducing individuals to data points rather than whole persons. When thoughtfully implemented, AI can complement rather than replace the relational aspects of nursing practice [9].

Artificial intelligence is increasingly being incorporated across multiple facets of healthcare, including nursing education and practice. Understanding nursing students’ perspectives on AI helps identify potential barriers to adoption and informs strategies for integrating AI effectively into nursing curricula and clinical environments [10]. Furthermore, examining these perspectives in Egypt, a country with a large and growing nursing workforce, may yield contextually grounded insights into the cultural and educational factors shaping AI adoption in low- and middle-income countries, thereby contributing to a more globally inclusive evidence base on AI in nursing.

In recent years, Egypt has increased national attention to artificial intelligence and digital transformation across multiple sectors, including healthcare and education. The government has adopted a more assertive approach to stimulate AI growth by promoting domestic research and development [11]. However, empirical evidence regarding nursing students’ knowledge and perceptions of artificial intelligence remains limited in the Egyptian educational context. The current study focuses primarily on general and clinical applications of AI, including machine learning algorithms, predictive analytics, and clinical decision-support systems, rather than exclusively on generative AI tools. These technologies analyse large volumes of data and enhance educational and training activities for nursing students.

Aim of the study

The study aimed to assess Egyptian nursing students’ knowledge of artificial intelligence, their perceptions of its benefits and fears regarding AI in nursing education and practice, and to examine associated sociodemographic factors.

Research questions

Q1

What is the level of AI knowledge among Egyptian nursing students, and what are their perceptions of AI’s benefits and risks in nursing education and practice?

Q2

What demographic factors (age, gender, academic year, university) are associated with nursing students’ AI knowledge, perceived benefits, and fears?

Theoretical framework

This study was conceptually informed by the Technology Acceptance Model (TAM) [12] and Rogers’ Diffusion of Innovations (DOI) Theory [13] as interpretive frameworks for understanding nursing students’ perceptions of artificial intelligence in nursing education and practice. TAM proposes that technology acceptance is influenced primarily by perceived usefulness and perceived ease of use, whereas DOI highlights how perceived relative advantage, compatibility, and complexity shape responses to innovation. The conceptual integration of these two frameworks, as applied in the present study, is illustrated in Fig. 1. In the present study, these frameworks were not formally tested; rather, they were used to help interpret patterns in students’ knowledge, perceived benefits, and fears regarding AI. Within this interpretive framing, perceived benefits may be broadly related to perceived usefulness and relative advantage, while fears may reflect concerns about complexity, compatibility, and ethical uncertainty. Accordingly, TAM and DOI served as sensitising perspectives that supported the interpretation of students’ responses without implying causal or predictive relationships.

Fig. 1.

Conceptual integration of the Technology Acceptance Model and Diffusion of Innovations Theory as interpretive frameworks for understanding nursing students’ perceptions of artificial intelligence

Methods

Study design

This study employed a quantitative, cross-sectional design to examine Egyptian nursing students’ perspectives on artificial intelligence and to assess relationships among AI knowledge, perceived benefits, and risks. The research adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist [14], ensuring transparent and methodologically rigorous reporting of this multi-university investigation.

Setting and sampling

This multi-university descriptive cross-sectional study was conducted at two Egyptian universities: Cairo University and Kafr Elsheikh University. The design was selected to provide a contemporaneous assessment of undergraduate nursing students’ knowledge of artificial intelligence and their perceptions of its benefits and risks in nursing education and practice across the two participating institutions.

A convenience sample of 2,412 undergraduate nursing students was recruited from different academic years. Eligible participants were students of both genders aged 18 to 24 years who were enrolled during the study period. Students who were absent at the time of data collection or declined participation were excluded.

Because the study aimed to estimate the mean AI knowledge score with a specified level of precision, the required sample size was calculated using the standard formula for estimating a population mean: n = (Z₁₋α/₂ × SD / d)², where Z₁₋α/₂ = 1.96 for a 95% confidence level, SD = 1.8 based on a comparable published study among nursing students assessing AI knowledge [15], and d = 0.07 as the prespecified margin of error. Substituting these values yielded a minimum required sample size of approximately 2,400 students. To improve the precision and stability of the estimates and to allow for potentially incomplete responses, a total of 2,412 students were included in the final analysis.

Data collection

The study instruments were selected following a comprehensive review of relevant Arabic and English literature on artificial intelligence in nursing and healthcare education. Written ethical approval was obtained from the ethics committee for scientific research at the Faculty of Nursing, Kafr Elsheikh University (KFSIRB200-243). In addition, official permission was obtained from the deans of the nursing faculties at Kafr Elsheikh University and Cairo University to conduct the study.

Three experts in the psychiatric and medical-surgical field assessed the content validity of the study tools. The data were collected using a structured questionnaire that included sections on knowledge, benefits, and fears regarding AI in nursing education. It was prepared to be completed in 10–15 min. The researchers explained the purpose of the study, assured participants of anonymity and confidentiality, and invited them to participate voluntarily. Students were also informed of their right to withdraw from the study at any time without penalty. Data were collected using Google Forms from May 2024 to January 2025.

Data collection tools

To ensure cultural and linguistic appropriateness for the Egyptian nursing student population, all study instruments were rigorously translated from English to Arabic using back-translation methodology [16, 17]. Two bilingual nursing education experts independently translated the original English scales into Arabic. These translations were then reconciled with the research team to produce a consensus Arabic version. A certified professional translator, blinded to the original English version, subsequently back-translated this into English. The research team, including all translators, systematically compared the original, translated, and back-translated versions to resolve any semantic discrepancies and finalise the Arabic instruments.

Before the main study, we conducted a pilot test with 30 nursing students (who were excluded from the final sample) to evaluate the clarity and comprehensibility of the translated measures. Participants confirmed the instruments were clear and culturally appropriate, requiring no substantive modifications. This comprehensive translation and validation process ensured the conceptual equivalence and psychometric integrity of all measures for our target population.

Part I: Self-administered survey questionnaire

Covered individual characteristics of student nurses like age, gender, and academic year.

Part 2: Artificial intelligence Knowledge questionnaire

Adapted from [17], it consisted of 5 multiple-choice items. One point was awarded for a correct answer, and zero for an incorrect one. Knowledge scores ranged from 0 to 5 and were categorised as: poor (0–1), considered less than 50%; average (2–3), considered 50%- 75%; and good (4–5), considered more than 75%.

Part 3: Questionnaire regarding perspective of benefits and fears

It was adapted from studies [18, 19] and structured into two sections: the first on the expected benefits of AI in academic education (7 questions) and the second on students’ fears toward AI (5 questions). Responses to closed questions were collected on a 3-point Likert scale from 0 to 2. All items were required to be completed for inclusion in the final analysis.

Validity and reliability

The study instruments showed satisfactory evidence of reliability and content adequacy. For the AI Knowledge scale, which comprised five dichotomous (correct/incorrect) items, internal consistency was assessed using the Kuder–Richardson Formula 20, yielding a coefficient of 0.896. For the Likert-scaled measures, Cronbach’s alpha demonstrated strong internal consistency for both the AI Benefits scale (α = 0.903) and the AI Fears scale (α = 0.899). All reliability coefficients exceeded the commonly accepted threshold of 0.70 [20].

Content validity was established through expert review. Three nursing education specialists and two researchers with expertise in artificial intelligence in healthcare evaluated the instruments for item relevance, clarity, and coverage of the study constructs. Based on their feedback, the items were deemed appropriate for assessing nursing students’ knowledge and perceptions of AI.

To further examine the internal structure of the adapted measures, confirmatory factor analysis was conducted in AMOS 26.0 using data from 2,412 nursing students. The hypothesised model specified the AI Knowledge scale (5 items), AI Benefits scale (7 items), and AI Fears scale (5 items) as distinct single-factor measures. The model showed acceptable overall fit to the data (χ²/df = 2.85, Comparative Fit Index = 0.95, Tucker–Lewis Index = 0.94, Root Mean Square Error of Approximation = 0.056, and Standardised Root Mean Residual = 0.042). Standardised factor loadings were statistically significant, ranging from 0.67 to 0.88. These findings provide supportive, though preliminary, evidence for the internal structure of the adapted instruments in this study sample.

Ethical consideration

Primary permission was secured from the research ethics committee at the Faculty of Nursing- Kafr Elsheikh University (KFSIRB200-243). Electronic informed consent (e-consent) was obtained from each participating student at the start of the online Google Forms questionnaire. The consent form clearly explained the study purpose, the voluntary nature of participation, the right to withdraw at any time without penalty, and assurances of anonymity and confidentiality; only students who actively confirmed agreement could proceed to the survey. The researchers emphasised that participation in the study is entirely voluntary. Confidentiality and anonymity were also assured.

Statistical analysis

Data were analysed using IBM SPSS Statistics version 26.0. Descriptive statistics, including frequencies, percentages, means, and standard deviations, were used to summarise participants’ demographic characteristics and questionnaire responses. Inferential statistics were applied to examine differences in AI knowledge, perceived benefits, and fears across demographic groups. Independent-samples t-tests were used for comparisons between two groups, and one-way analysis of variance (ANOVA) was used for comparisons across more than two groups. Post-hoc pairwise comparisons were performed using Tukey’s honestly significant difference (HSD) test, where ANOVA results were significant. Statistical significance was set at p < 0.05.

To reduce the risk of response bias, several procedural steps were implemented during questionnaire development and administration. Survey items were reviewed by experts and pretested to improve clarity and reduce ambiguity. Participants received standardised instructions emphasising that there were no right or wrong answers and that responses would be used only for research purposes. Anonymity was maintained throughout data collection, and no identifying information was collected. These procedures were intended to encourage candid responses and reduce the likelihood of socially desirable or patterned responding [21].

Results

As shown in Table 1, the mean age of the participating nursing students was 20.64 ± 1.3 years. Students younger than 21 years accounted for 52.6% of the sample, whereas 47.4% were aged 21–24 years. Most participants were female (74.2%), while males represented 25.8% of the sample. Regarding academic year, 2.8% were first-year students, 22.2% were second-year students, 30.3% were third-year students, and 44.6% were fourth-year students. The sample was drawn almost equally from the two participating universities, with 48.3% from Cairo University and 51.7% from Kafr Elsheikh University.

Table 1.

Distribution of the demographic characteristics of the study sample (N = 2412)

Student characteristics	N	%
Age
< 21	1268	52.6
21–24	1144	47.4
Mean ± SD	20.64 ± 1.3
Gender
Male	623	25.8
Female	1789	74.2
Academic year
1st	68	2.8
2nd	536	22.2
3rd	732	30.3
4th	1076	44.6
University
Cairo University	1164	48.3
Kafr Elsheikh University	1248	51.7

Table 2 shows that 87.9% of students correctly identified the meaning of artificial intelligence, and 79.9% reported having heard about AI. In addition, 68.4% correctly recognised uses of AI in the health sector. However, 62.3% reported not receiving any formal AI-related training as part of their curriculum. Overall, 46.4% of participants demonstrated good AI knowledge, 44.9% demonstrated average knowledge, and 8.7% demonstrated poor knowledge, with a mean knowledge score of 3.2 ± 1.3 out of 5.

Table 2.

Distribution of the Knowledge level among the study sample (N = 2412)

Items	Incorrect		Correct
	N	%	n	%
Meaning of artificial intelligence	292	12.1	2120	87.9
Knowledge of AI applications in the health sector	762	31.6	1650	68.4
Awareness that AI is widely used in the country	1332	55.2	1080	44.8
Hearing about artificial intelligence	486	20.1	1926	79.9
Received any formal AI-related training within the curriculum	1502	62.3	910	37.7
Overall knowledge level	N		%
Poor knowledge	209		8.7
Average knowledge	1084		44.9
Good knowledge	1119		46.4
Mean ± SD	3.2 ± 1.3

As shown in Table 3, students generally reported favourable perceptions of AI’s benefits. More than three-quarters agreed that AI can facilitate the gathering and analysis of patient information (76.8%), assist in identifying nursing diagnoses (76.8%), and support medical and nursing research (77.9%). Similarly, 76.5% agreed that AI can promote healthier patient behaviours, and 74.7% agreed that it can improve patient compliance and follow-up. By contrast, agreement was lower for the item stating that AI improves nursing training (38.2%). Overall, the mean perceived benefits score was 10.9 ± 3.5.

Table 3.

Distribution of perceived benefits of AI in nursing education among the study sample (N = 2412)

Perceived benefits of AI in nursing education	No		Neutral		Yes
	n	%	n	%	n	%
AI helps improve nursing training	576	23.9	914	37.9	922	38.2
AI facilitates gathering and analysing patient information	202	8.4	358	14.8	1852	76.8
AI assists in identifying nursing diagnoses	32	1.3	528	21.9	1852	76.8
AI can support medical and nursing research	282	11.7	250	10.4	1880	77.9
AI can support healthier patient behaviours	352	14.6	216	9.0	1844	76.5
AI can improve patient adherence and follow-up	40	1.7	570	23.6	1802	74.7
AI-related education may help address ethical challenges in practice	474	19.7	842	34.9	1096	45.4
Mean ± SD	10.9 ± 3.5

Table 4 shows that fears regarding AI were also prominent. The most commonly reported concern was that AI could replace nurses in the future, endorsed by 72.2% of students. In addition, 57.8% reported discomfort with the use of AI in educational settings, and 55.4% viewed AI as a threat to patient data confidentiality. The mean fear score was 5.0 ± 1.6. Figure 2 illustrates students’ perceptions of key challenges related to AI use in healthcare. Additional concerns included the belief that AI cannot be used in the care of human beings (39.3%), limited concern about AI’s impact on emotional well-being (36.2%), and the perception that AI is not sufficiently flexible for individual patient needs (47.4%).

Table 4.

Distribution of nursing students’ fears regarding future AI use (N = 2412)

Fears regarding AI	No		Neutral		Yes
	n	%	n	%	n	%
AI may reduce future employment opportunities in healthcare	882	36.6	776	32.2	754	31.3
AI may replace nurses in the future healthcare system	514	21.3	156	6.5	1742	72.2
Comfortable using AI in health education (No = discomfort)	1394	57.8	252	10.4	766	31.8
AI threatens the human dimension of care	1472	61.0	154	6.4	786	32.6
AI threatens patient data confidentiality	1076	44.6	0	0.0	1336	55.4
Mean ± SD	5.0 ± 1.6

Fig. 2.

Distribution of student nurses regarding problems of the application of AI in health

Table 5 shows statistically significant differences across age groups for all three AI-related outcomes. Compared with students younger than 21 years, those aged 21–24 years reported significantly higher perceived benefits and significantly lower fear toward AI (t = 5.805, p < 0.001; t = 14.491, p < 0.001, respectively), whereas students younger than 21 years had a slightly higher mean knowledge score (t = 2.664, p = 0.008). Academic year also showed statistically significant differences in knowledge, perceived benefits, and fears (all p < 0.001). Fourth-year students had the highest mean perceived benefits score and the lowest mean fear score, whereas first-year students had the lowest mean knowledge score and the highest mean fear score. No statistically significant differences were observed by gender or university.

Table 5.

Differences in AI knowledge, perceived benefits, and fears according to demographic characteristics among nursing students (N = 2412)

Demographic characteristics	AI knowledge Mean ± SD	Perceived benefits Mean ± SD	Fears Mean ± SD
Age
< 21 years	3.3 ± 1.4	10.5 ± 3.6	5.5 ± 1.3
21–24 years	3.1 ± 1.2	11.3 ± 3.4	4.5 ± 1.8
Independent-samples t-test	t = 2.664, p = 0.008	t = 5.805, p < 0.001	t = 14.491, p < 0.001
Gender
Male	3.1 ± 1.3	10.8 ± 3.5	5.0 ± 1.7
Female	3.2 ± 1.3	10.9 ± 3.5	5.1 ± 1.6
Independent-samples t-test	t = 0.541, p = 0.589	t = 0.113, p = 0.910	t = 0.052, p = 0.958
Academic year
1st year	1.9 ± 0.9	7.2 ± 2.4	7.5 ± 1.2
2nd year	3.1 ± 1.1	10.3 ± 2.8	5.4 ± 1.3
3rd year	3.3 ± 1.6	10.7 ± 3.9	5.5 ± 1.4
4th year	3.2 ± 1.1	11.7 ± 3.1	4.4 ± 1.7
One-way ANOVA	F = 24.103, p < 0.001	F = 127.881, p < 0.001	F = 158.804, p < 0.001
University
Cairo University	3.1 ± 1.3	10.9 ± 3.5	5.1 ± 1.6
Kafr Elsheikh University	3.1 ± 1.4	10.8 ± 3.6	5.0 ± 1.6
Independent-samples t-test	t = 0.088, p = 0.930	t = 0.171, p = 0.864	t = 0.018, p = 0.986

Discussion

Principal findings

This multi-university cross-sectional study provides a nuanced picture of Egyptian nursing students’ perspectives on artificial intelligence in nursing education and practice. Three findings are especially salient. First, students demonstrated a foundational but uneven level of AI knowledge, with 46.4% showing good knowledge and 44.9% showing average knowledge. Second, students broadly recognised several benefits of AI, particularly in relation to information gathering and analysis, support for identifying nursing diagnoses, assistance with medical and nursing research, and improvements in treatment follow-up and adherence. Third, these positive perceptions coexisted with substantial fears, particularly concern that AI may replace nurses, discomfort with AI use in educational settings, and worry about patient data confidentiality. Collectively, these findings suggest not rejection of AI, but a cautious and conditional openness shaped by partial knowledge, limited formal preparation, and unresolved professional and ethical concern [22, 31, 32, 36, 37].

Interpretation of nursing students’ AI knowledge, perceived benefits, and fears

A central finding of this study is the mismatch between general awareness of AI and formal educational exposure to it. Most students had heard of AI and correctly understood its meaning, yet 62.3% reported not receiving specialised AI-related training in their curriculum. This indicates that students may be forming perceptions of AI from informal, fragmented, or media-based exposure rather than from structured nursing education. A similar pattern has been observed in prior work, which shows growing awareness of AI among health-related students and professional groups despite gaps in formal preparation [22, 23]. This curricular gap is educationally important because superficial awareness, in the absence of guided learning, may foster uncertainty rather than competence.

The pattern of perceived benefits in this study was also instructive. Students most strongly endorsed AI when it was framed as a supportive tool for gathering and analysing patient information, assisting with nursing diagnosis identification, and supporting medical and nursing research, findings that align with prior literature emphasising AI’s strengths in data processing, decision support, and educational assistance [24, 25]. Students also recognised AI’s potential role in promoting healthier patient behaviour and improving treatment adherence and follow-up, consistent with evidence that AI-supported personalisation may strengthen patient engagement and continuity of care [26, 27]. By contrast, endorsement was more moderate for the broader claim that AI improves nursing training, suggesting that students may be more persuaded by concrete and bounded applications than by general claims about educational transformation [1, 9, 33, 34].

The ethical dimension of the findings requires careful attention. Although 45.4% agreed with the item concerning ethical issues and AI, more than half also viewed AI as a threat to patient data confidentiality [28, 29]. Taken together, these responses suggest that students recognise ethical vulnerability but may not yet have a sufficiently clear framework for understanding how ethical AI training should be operationalised in nursing education. This interpretation is consistent with previous literature emphasising the need for transparency, accountability, and robust data governance in healthcare AI [30, 31].

The findings on fear were among the most revealing in the study. The strongest concern was that AI would replace nurses in the future, yet this was not matched by the lower proportion who feared losing their jobs. Rather than representing a contradiction, this divergence suggests a distinction between symbolic concern about professional role redefinition and direct anxiety about personal employment loss. Students may therefore be responding to broader narratives about automation in healthcare while still recognising that nursing involves relational, ethical, and judgment-based dimensions that are not easily displaced by AI. This interpretation is also consistent with prior work on technology anxiety and professional uncertainty in health professions education [32, 33].

Discomfort with AI use in educational settings and concern about patient data confidentiality reinforce the point that students’ concerns are not merely abstract. They reflect practical unease about trust, safety, and appropriateness. Likewise, the finding that many students believed AI may lack the flexibility to address the unique needs of individual patients suggests that students are already evaluating AI through a person-centred nursing lens [34]. This concern aligns with prior evidence indicating scepticism among healthcare workers regarding AI’s capacity to accommodate individual patient complexity [35]. Rather than dismissing these fears, nursing education should engage them directly through structured teaching on safe use, boundaries of application, and the continuing centrality of human clinical judgment [36].

Theoretical interpretation through the Technology Acceptance Model and the Diffusion of Innovations theory

The findings may be interpreted through the Technology Acceptance Model and Diffusion of Innovations theory as broad conceptual lenses rather than as formally tested explanatory models. From a TAM perspective, the benefits items most closely align with perceived usefulness, as students were more likely to endorse AI when its value for information processing, research support, and clinical assistance was evident [12]. Although perceived ease of use was not directly measured, the coexistence of favourable perceptions of benefits and substantial fears suggests that students may view AI as potentially useful while remaining uncertain about its practical, ethical, or professional implications.

From the TAM perspective specifically, greater foundational knowledge of AI may enhance perceived ease of use by making the technology appear less complex and more accessible. In this regard, formal training may be understood as an important educational condition shaping students’ responses to AI, even though it was not tested as a statistical moderator in the present study. Structured learning can reduce uncertainty, clarify practical applications, and make AI systems appear more intelligible and manageable. From the DOI perspective, this is relevant to the complexity attribute because innovations perceived as difficult to understand or use are more likely to generate resistance. Interpreted together, the TAM and DOI frameworks suggest that the relative lack of formal AI training reported by many students may help explain why AI remained associated with fear and uncertainty, whereas greater educational exposure may reduce perceived complexity and thereby enhance perceived ease of use. The finding that students in higher academic years reported lower fear and greater perceived benefits is consistent with this interpretation, although the cross-sectional design does not permit causal inference.

The DOI framework offers a complementary interpretation. Strong endorsement of selected benefits may reflect perceived relative advantage, whereas fear-related responses may reflect perceived complexity and concerns about compatibility with nursing values and person-centred care [13]. The more favourable pattern observed among students in higher academic years is consistent with the possibility that greater exposure to professional and educational contexts may be associated with more positive perceptions of AI. However, given the cross-sectional design, these patterns should be interpreted as associations rather than causal relationships [9].

Implications for nursing education and practice

The findings carry direct implications for undergraduate nursing education. First, AI content should be introduced earlier and more systematically in the curriculum rather than left to informal exposure. Second, AI education should move beyond technical description to include supervised discussion of ethics, data confidentiality, accountability, bias, and the limits of AI in individualised care. Third, educators should explicitly distinguish role augmentation from role replacement. In the present study, students’ concerns appeared to reflect not only employment-related anxiety but also deeper unease about professional role redefinition and the possible loss of the human, relational, and judgment-based dimensions of nursing practice; educational discussion should therefore address these concerns directly rather than treating them simply as fear of technological substitution. Addressing this distinction directly may help reduce unnecessary anxiety and foster a more informed, professionally grounded acceptance of AI. Finally, practical teaching approaches such as case-based learning, simulation, and guided critique of AI-supported outputs may help students develop not only digital literacy but also the professional judgment needed to use AI safely and responsibly in nursing contexts [24, 34, 37].

Conclusion

This multi-university cross-sectional study shows that Egyptian nursing students hold a cautiously receptive view of artificial intelligence in nursing education. Students demonstrated foundational AI awareness and recognised several perceived benefits, particularly in information gathering and analysis, support for nursing diagnosis identification, research assistance, and treatment follow-up. However, these favourable perceptions coexisted with substantial gaps in formal AI training and persistent concerns, especially fear of nurse replacement, discomfort with AI use in educational settings, and worries about data confidentiality. The significant differences observed across academic year and age suggest that students’ AI-related perceptions may vary with educational stage and level of exposure. These findings support the integration of structured, ethically grounded AI education into undergraduate nursing curricula, with emphasis on practical competence, data privacy, accountability, and the evolving professional role of nurses in AI-enabled care. Future research should examine whether curriculum-based AI education improves students’ competence, ethical readiness, and preparedness for technology-enabled nursing practice.

Limitations

Several limitations should be considered when interpreting these findings. First, the cross-sectional design precludes causal inference and does not capture how students’ perceptions may change over time. Second, the relatively extended data collection period (May 2024 to January 2025), together with the rapidly evolving nature of AI technologies and public discourse, may have introduced temporal variation in participants’ views. Third, although rigorous translation and validation procedures were undertaken, the instruments were adapted rather than originally developed for nursing education contexts, which may limit measurement specificity. Fourth, the use of self-reported data may have introduced social desirability and other response biases, despite the procedural steps taken to reduce these risks during questionnaire development and administration. Fifth, convenience sampling from only two Egyptian universities may limit representativeness and reduce generalizability to other educational or national contexts. Sixth, potentially important cultural influences on students’ perceptions of AI were not directly measured. Seventh, the first-year subgroup was substantially smaller than the other academic-year groups; therefore, findings for this subgroup should be interpreted with caution. Eighth, both the ‘AI will replace nurses’ and ‘risk of losing jobs’ items were retained verbatim from the adapted source instrument [18]. Their divergent endorsement rates (72.2% vs. 31.3%) were interpreted as reflecting different levels of conceptual versus personal employment concern; however, this distinction was inferred rather than directly measured and should therefore be interpreted with caution. Future studies should use probability-based and multi-site sampling, include objective assessments of AI competence, and adopt longitudinal or interventional designs to evaluate whether structured AI education improves preparedness and reduces fear.

Abbreviations

AI

Artificial Intelligence

AMOS

Analysis of Moment Structures

ANOVA

Analysis of Variance

DOI

Diffusion of Innovations

HSD

Honestly Significant Difference

IBM

International Business Machines

KR-20

Kuder–Richardson Formula 20

ML

Machine Learning

SD

Standard Deviation

SPSS

Statistical Package for the Social Sciences

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

TAM

Technology Acceptance Model

Author contributions

NAM contributed to the study conception and design, data collection, data curation, interpretation of findings, manuscript drafting, and critical revision. SOA contributed to the study design, instrument preparation, interpretation of findings, and critical revision of the manuscript. OMER contributed to manuscript development, methodological refinement, interpretation of findings, critical revision, and preparation of the manuscript for submission. SMF contributed to data collection, data analysis, interpretation of findings, and manuscript revision. EEEE contributed to instrument preparation, data collection, interpretation of findings, and critical review of the manuscript. MEZ contributed to statistical analysis, interpretation of findings, manuscript revision, and final review. All authors read and approved the final manuscript.

Funding

This study is supported via funding from Prince Sattam Bin Abdulaziz University project number (PSAU/2026/R/1447).

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was obtained from the Nursing Research Ethics Committee, Faculty of Nursing, Kafr Elsheikh University, Egypt (KFSIRB200-243). Electronic informed consent was obtained from all participants before completing the questionnaire. Participation was voluntary, and all procedures were conducted in accordance with the Declaration of Helsinki. Confidentiality and anonymity were maintained throughout the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.