The Impact of Artificial Intelligence Attitudes and Acceptance on Critical Thinking Motivation among Nursing Students in Saudi Arabia

Abstract

Introduction

As Saudi Arabia advances its Vision 2030 agenda, which emphasizes artificial intelligence (AI) integration in healthcare and education, understanding students’ acceptance of AI in academic settings is increasingly important.

Objective

To examine the relationship between nursing students’ attitudes toward AI, perceived usefulness, perceived usability, and their critical thinking motivation.

Methods

A multicenter cross-sectional correlational study was conducted across three Saudi universities. Data were collected from 390 undergraduate nursing students using the General Attitudes towards AI Scale, the technology acceptance model questionnaire, and the critical thinking motivation scale. Analysis included Pearson correlations and multiple linear regression.

Results

Students reported positive attitudes toward AI, with strong correlations among AI attitude, usefulness, and usability (r = .63–.78, p < .001). Weak but significant positive correlations were found between AI-related factors and critical thinking motivation (r = .20–.40, p < .001). Higher academic level, the belief that AI will not replace nursing roles, and greater AI usability significantly predicted critical thinking motivation (adjusted R² = .245, p < .001).

Conclusion

Acceptance and perceived usability of AI tools are positively associated with critical thinking motivation in nursing students. These findings underscore the potential of AI to support cognitive skill development in nursing education. Educators should integrate AI early in nursing curricula, provide structured training, and frame AI as a support tool in clinical decision-making. Doing so can foster both critical thinking and digital competence among future nurses.

Introduction

Globally, artificial intelligence (AI) is increasingly recognized as a transformative tool in healthcare, prompting efforts to integrate AI into nursing education. Many nursing programs are beginning to incorporate AI-driven technologies—such as virtual simulations, intelligent tutoring systems, and chatbots—to enhance learning and prepare students for a technology-rich clinical environment (Labrague et al., 2023; Zhang et al., 2024). By leveraging AI in education, these initiatives aim to improve students’ competencies in data-driven decision-making and familiarize future nurses with tools they may encounter in practice. However, the pace of integration remains uneven worldwide, and the success of such innovations largely depends on learners’ acceptance and understanding of AI's relevance to patient care. Nursing students’ perceptions of AI, therefore, have become a crucial focus, as positive perceptions are believed to facilitate the successful adoption of AI in their future clinical practice (Labrague et al., 2023).

As part of this global trend, Saudi Arabia has prioritized digital transformation in both education and healthcare, driven by the National Vision 2030 Initiative. Vision 2030 explicitly calls for integrating advanced technologies such as AI to modernize society and workforce skills (Saudi Arabian Government, 2016). In line with this strategic vision, the country has launched a national AI strategy and invested in AI infrastructure and training programs (Housawi & Lytras, 2023). These top-down efforts have started to permeate nursing education: institutions are exploring AI applications ranging from e-learning platforms to clinical simulation tools, reflecting a mandate to leverage technology for educational excellence. Initial evidence suggests that Saudi nursing students are aware of AI's growing role and are generally optimistic about its integration. For instance, in one regional survey, 79% of nursing students supported adding AI training to the nursing curriculum (Salama et al., 2025). At the same time, many students reported having little formal exposure to AI in their programs (Salama et al., 2025), highlighting a gap between Vision 2030's digital ambitions and the current state of nursing curricula.

In professional practice, nurses are increasingly encountering AI applications across diverse clinical settings. AI is being used to support decision-making in patient triage, predict patient deterioration using early warning scores, assist with medication administration through smart infusion systems, and streamline documentation through natural language processing tools (Amin et al., 2024; Zhang et al., 2024). For instance, clinical decision support systems powered by AI can analyze vast patient data to flag high-risk cases, enabling nurses to prioritize care more effectively. In radiology departments, AI supports nurses in interpreting medical imaging by offering preliminary findings, while wearable technologies enable continuous monitoring of patient vitals, reducing the need for constant bedside presence and enhancing early intervention protocols. Moreover, virtual nursing assistants are being piloted to provide real-time guidance during clinical procedures, reducing errors and enhancing workflow efficiency.

Despite these advancements, many nurses remain cautious about AI integration, citing concerns about data accuracy, ethical accountability, and potential job displacement (Alruwail et al., 2025; Amin et al., 2024). A lack of formal training and uncertainty about AI's reliability in high-stakes situations often contribute to resistance. These realities underscore a critical gap between technological innovation and workforce readiness, where attitudes toward AI—shaped during undergraduate education—may significantly influence its acceptance and effectiveness in practice. This context underscores the importance of assessing nursing students’ perception of AI within Saudi Arabia, as their readiness and acceptance will critically influence how effectively educational reforms involving AI can be implemented.

Parallel to the rise of AI, there is continued emphasis on cultivating critical thinking in nursing education. Effective clinical decision-making hinges on strong critical thinking skills, as nurses must analyze complex patient information, evaluate options, and make sound judgments to ensure safe care. Critical thinking is thus widely regarded as an essential competency for nursing practice (Papathanasiou et al., 2014). Nursing programs globally incorporate pedagogies such as problem-based learning, simulation, and reflective practice to strengthen students’ critical thinking abilities (Al-Otaibi et al., 2023; Papathanasiou et al., 2014).

Yet, beyond teaching cognitive skills, educators recognize that students’ motivation to engage in critical thinking is equally crucial. Motivational factors often determine whether learners actively apply critical reasoning in practice or revert to rote decision-making. Recent research in nursing education has introduced the notion of critical thinking motivation (CTM)—the drive or willingness to employ critical thinking skills in learning and clinical situations. For example, using motivational teaching strategies (e.g., interactive and student-centered activities) has been shown to increase nursing students’ engagement in critical thinking processes (Al-Otaibi et al., 2023). In other words, students who are more motivated and see personal value in analytical problem-solving tend to delve deeper into case analysis and demonstrate greater initiative in clinical learning. This suggests that fostering an environment that not only teaches critical thinking skills but also inspires students to use those skills is vital for effective nursing education.

Review of Literature

This study draws on two complementary theoretical frameworks: the technology acceptance model (TAM) and self-determination theory (SDT). Both models have been extensively applied in healthcare and education to examine users’ engagement with technology and learning behaviors.

TAM, developed by Davis (1989), posits that an individual's acceptance and use of technology are primarily driven by two constructs: perceived usefulness (the belief that using the system will enhance performance) and perceived ease of use (the belief that the system is free of effort). These constructions have been validated in healthcare contexts, including among nursing students and educators (Labrague et al., 2023; Salama et al., 2025). In this study, TAM guided the selection of key independent variables: general attitudes toward AI, perceived usefulness, and perceived usability (ease of use), all of which were hypothesized to shape students’ willingness to engage with AI in their academic and clinical training.

SDT, introduced by Ryan and Deci (2000), explains how human motivation is influenced by the fulfillment of three basic psychological needs: autonomy, competence, and relatedness. When learners feel competent in using a tool, autonomous in applying it, and connected to the learning process, they are more likely to exhibit intrinsic motivation, which is essential for higher-order thinking skills such as analysis, evaluation, and decision-making (Jia & Tu, 2024; Papathanasiou et al., 2014).

By integrating TAM and SDT, this study conceptualizes CTM not simply as an outcome of individual traits but as a contextual response to technological and environmental factors. We hypothesized that positive perceptions of AI (TAM) would reinforce students’ sense of competence and autonomy (SDT), thereby enhancing their motivation to engage in critical thinking. This framework also informed our interpretation of results: for example, the finding that AI usability significantly predicted CTM is consistent with both TAM (ease of use leading to greater adoption) and SDT (tools that feel manageable enhance intrinsic engagement).

Existing studies have focused on attitudes or knowledge levels without linking these constructs to motivation. Research from the Middle East, specifically Saudi Arabia, remains sparse and lacks theoretical integration. This study addresses this gap by examining the interplay between AI attitudes, perceived usefulness, and CTM through TAM and SDT. This dual-framework approach is particularly relevant in Saudi Arabia, where students are increasingly exposed to AI technologies but may lack structured training (Alatawi et al., 2024; Al-Otaibi et al., 2023). In such environments, perceptions of technology may significantly shape motivation, especially if students view AI tools as empowering or undermining their autonomy. By bridging TAM's focus on technology engagement with SDT's emphasis on psychological motivation, this study offers a holistic understanding of AI readiness and CTM in evolving educational systems.

Objectives

Based on the preceding literature and theoretical considerations, this study pursues the following objectives:

• To explore the relationship between the acceptance, attitudes toward AI, and CTM among nursing students in Saudi Arabia.

• To identify significant factors affecting CTM among Suaidian nursing students

Methods

Design

This study employs a multicenter cross-sectional design aligned with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. A cross-sectional approach is suitable as it captures a snapshot of the participants’ attitudes, perceptions, and motivational status at a specific point in time (Krejcie & Morgan, 1970). The data collection for this study spanned three months, from September 2024 to December 2024.

Setting

The study was conducted across three universities in Saudi Arabia, selected for their strong nursing programs, geographic spread, and institutional cooperation. These sites included the nursing college at Al-Qassim University, a large public university in the Qassim region (central Saudi Arabia), along with the nursing college at Jouf University and the nursing college at Imam Abdulrahman Bin Faisal University. Moreover, all three universities demonstrated a willingness to participate in and facilitate the research, providing the necessary administrative support and access to nursing students. The research was conducted over three months, starting at the beginning of September 2024 and concluding at the end of December 2024.

Sample

The study initially calculated a minimum sample size of 338 participants using the Epi-Info program based on a population of 1,200 nursing students. The calculation considered an expected proportion of 50%, an acceptable margin of error of 5%, a design effect of 1, a confidence level of 97%, and a statistical power of 80%. To mitigate the risk of attrition bias and nonresponses, the sample size was increased to 390 participants. This adjustment ensured that the final sample, after accounting for potential dropouts and incomplete responses, would still maintain the necessary statistical power and confidence. The total number of enrolled students was 118 for the nursing college at Imam Abdulrahman Bin Faisal University, 172 for the college of nursing at Al-Qassim University, and 100 for the nursing college at Jouf University (total number 390).

Inclusion Criteria

Undergraduate nursing students in a bachelor's degree program and willing to participate and complete the survey.

Exclusion Criteria

Students enrolled in bridging programs (e.g., diploma-to-BSN bridging courses) or postgraduate nursing programs (e.g., master's or diploma specializations) were excluded, and responses were incomplete or inconsistent.

Sampling Technique

A nonprobability convenience sampling method was employed. Participants were recruited based on their availability and willingness to participate in the survey at the selected universities. This technique was chosen due to the practical advantages of accessing respondents who were readily available through the universities’ nursing programs and classes. Given the cooperation of the institutions, researchers could conveniently reach many nursing students during scheduled academic activities.

Study Tools (see Supplemental File)

General Attitudes Towards AI Scale (GAAIS)

Developed by Schepman and Rodway in 2020, the GAAIS evaluates individuals’ positive and negative views on AI. It consists of 20 items split into two subscales: positive and negative attitudes. The positive subscale includes statements such as “I am interested in using AI in my daily life,” “There are many beneficial applications of AI,” and “AI can create new economic opportunities.” On the other hand, the negative subscale features statements such as “I believe AI is dangerous,” “Organizations could misuse AI,” and “The future uses of AI make me feel uneasy.” These items are scored using a 5-point Likert scale (from Strongly Agree = 5 to Strongly Disagree = 1), with the negative items being reverse-scored. The overall score ranges from 20 to 100, with higher scores indicating more favorable attitudes toward AI in healthcare. The scale has been shown to have strong psychometric properties, including good convergent and discriminant validity, and is cross-validated in various studies Cronbach's alpha ranging from 0.84 to 0.89 (Schepman & Rodway, 2020). The reliability of the scale in this study, measured by Cronbach's alpha, was 0.871.

TAM Questionnaire

The TAM questionnaire, originally developed by Davis in 1989 to assess users’ acceptance of information systems and technology, was adapted for this study to measure nursing students’ perceptions of perceived AI usability. The original “Technology” term was replaced with “AI.” The questionnaire comprises two subscales: “Perceived Usefulness” and “Perceived Ease of Use,” each containing six items. The responses are scored on a 5-point Likert scale (ranging from Strongly Disagree = 1 to Strongly Agree = 5), with total scores ranging from 12 to 60. Higher scores reflect more favorable perceptions of AI usability in healthcare among nursing students. Prior studies in healthcare and educational contexts have reported reliability coefficients between 0.78 and 0.88 for the perceived usefulness and ease-of-use subscales (Davis, 1989; Labrague et al., 2023). In this study, Cronbach's alpha reliability estimate was 0.810.

CTM Scale (CTMS)

Developed by Valenzuela et al. in 2011, the CTMS is grounded in the theoretical perspective that motivation plays a key role in fostering critical thinking, emphasizing its importance over attitudes. The scale aims to assess students’ levels of motivation and identify the specific components that influence their performance. The CTMS consists of 19 items on a 6-point Likert scale, measuring various motivational aspects related to critical thinking. It examines participants’ expectations (expectation) and the perceived importance (value) of thinking critically, along with the utility they find in such thinking, the costs they are willing to pay, and the interest this type of thinking generates. Valenzuela et al. (2011) originally reported Cronbach's alpha of 0.77, consistent with our finding of 0.912.

In this study, all three instruments (GAAIS, TAM questionnaire, and CTMS) were administered in English. This choice was consistent with the standard practice in Saudi nursing programs, where English is commonly used as the medium of instruction.

Ethical Consideration

In accordance with the World Medical Association's Code of Ethics outlined in the Declaration of Helsinki (DoH 2008), approval for this study was granted by the Research Ethical Committee (REC) at Imam Abdulrahman Bin Faisal University's Faculty of Nursing, Saudi Arabia (IRB-2024-04-621) granted authorization for data collection in this study on 15 September 2024. Rigorous procedures were implemented to protect the confidentiality and privacy of all data collected. Prior to participation, all individuals were required to sign a written informed consent form. Participants were fully informed of the voluntary nature of their involvement and assured that withdrawing consent at any point would not result in any negative consequences.

Data Collection

Students were invited to participate through email announcements and notifications via the universities’ virtual learning management systems. The invitation provided detailed information about the study's objectives, the voluntary nature of participation, the confidentiality of responses, and the study's ethical guidelines. Students who expressed interest in participating were required to provide electronic informed consent.

Participants completed the survey via a secure, encrypted online platform. The online version used an encrypted survey link to ensure the confidentiality and security of the data collected. The survey was designed to be completed in approximately 15–20 min, and participants could complete it at their convenience within the study period.

No personal identifiers, such as names or student IDs, were collected to maintain participant anonymity and protect respondents’ privacy. All responses were stored securely on password-protected devices that were only accessible to the principal investigators. Any identifiable information was intentionally excluded to further safeguard participants’ privacy, adhering to the principles of confidentiality outlined in the ethical approval process. By employing these rigorous data collection methods, the study ensured both ethical integrity and the security of sensitive information while obtaining a comprehensive set of responses from the nursing students.

Statistical Analysis

Data were analyzed using IBM SPSS Statistics, Version 27 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize participants’ sociodemographic characteristics, including frequencies, percentages, means, and standard deviations.

To examine the relationships between variables, bivariate analyses were first conducted. These included independent sample t tests and one-way analysis of variance to compare CTM across categorical variables (e.g., academic level, AI tool used, and course attendance). Pearson's correlation coefficient was used to assess the strength and direction of linear relationships between continuous variables (e.g., AI attitudes, AI usefulness, AI usability, grade point average (GPA), and CTM).

Following the bivariate analysis, a multiple linear regression model was constructed to identify significant predictors of CTM. Only variables that demonstrated statistical significance (p < .05) in bivariate analyses or were theoretically relevant based on the conceptual framework (TAM and SDT) were included in the regression model. For example, GPA was retained due to its significant positive correlation with CTM. In contrast, AI course attendance, despite being significant in bivariate tests, was excluded from the final model due to concerns of multicollinearity and its weak effect size when adjusted for other variables. Variables such as gender, residence, and AI usage duration, which did not show significant bivariate associations, were excluded to maintain model parsimony.

Multicollinearity among predictor variables was assessed using the variance inflation factor, with a cutoff value of 5 to determine acceptability. No substantial multicollinearity was found among the final included variables. Assumptions of normality, linearity, homoscedasticity, and independence of residuals were checked through graphical methods (e.g., Q–Q plots and residual scatterplots) and statistical tests (e.g., Shapiro–Wilk test), and no violations were detected. The results of the regression analysis were reported using standardized beta coefficients (β), adjusted R², and F-statistic values to assess model fit and explanatory power. Statistical significance was set at p < .05 for all analyses.

Results

Table 1 illustrates the sociodemographic characteristics of the students. The majority of students were male (74.4%) and from urban areas (82.6%). The mean GPA was 4.30 ± 0.58. The majority of students were not working during their study (87.4%). The main source of information about AI was the internet (78.2%), and ChatGPT was the most used application (55.6%; Table 1).

Table 1.

Sociodemographic Characteristics of Students (N = 390).

Variables	Categories	n	%
Age	20.49 ± 1.20
Gender
	Male	290	74.4
	Female	100	25.6
Academic year
	First	22	5.6
	Second	143	36.7
	Third	147	37.7
	Fourth	78	20.0
Residence
	Urban	322	82.6
	Semiurban	36	9.2
	Rural	32	8.2
GPA	4.30 ± .58
Do you work during your study?
	Yes	49	12.6
	No	341	87.4
What are your sources of information about AI?
	Internet	305	78.2
	Friends	71	18.2
	Workshop	14	3.6
What are these applications?
	ChatGPT	217	55.6
	Deepseek	95	24.4
	Gemini	78	20.0
Hours use
	1.00	223	57.2
	2.00	74	19.0
	3.00	58	14.9
	4.00	24	6.2
	5.00	11	2.8
Period of using AI applications (in years) 1.41 ± .76
Did you attend training courses about AI usage in nursing previously?
	Yes	66	16.9
	No	324	83.1

Note. GPA = grade point average; AI = artificial intelligence.

As illustrated in Table 2, the means of AI attitudes, AI usefulness, and AI usability were 64.93 ± 8.56, 22.54 ± 4.96, and 21.96 ± 4.40, respectively. The mean of CTM was 86.12 ± 18.69, with a range of 95 (19–114).

Table 2.

Mean of Study Variables.

Variables	Range	Minimum	Maximum	Mean	Std. Deviation
AI attitudes	62	30	92	64.93	8.56
AI usefulness	24	6	30	22.54	4.96
AI usability	24	6	30	21.96	4.40
Critical thinking motivation	95	19	114	86.12	18.69

Note. AI = artificial intelligence.

As shown in Table 3, there was a strong positive correlation between AI attitudes and AI usefulness (r = .653, p < .001) and AI usability (r = .627, p < .001). In addition, a strong positive correlation was found between AI usefulness and AI usability (r = .787, p < .001). However, there was a weak positive correlation between critical care thinking and AI attitudes (r = .302, p < .001), AI usefulness (r = .348, p < .001), and AI usability (r = .397, p < .001).

Table 3.

Correlation Between Study Variables.

Variables		AI attitudes	AI usefulness	AI usability	CTM
AI attitudes	r	1
	p
AI usefulness	r	.653**
	p	<.001
AI usability	r	.627**	.787**
	p	<.001	<.001
Critical thinking motivation	r	.302**	.348**	.397**	1
	p	<.001	<.001	<.001

Note. AI = artificial intelligence; CTM = critical thinking motivation. **Correlation is significant at the 0.01 level (2-tailed).

Significant findings include variations in CTM based on academic year, with fourth-year students reporting the highest motivation scores (p < .001). Nonworking students exhibited significantly higher CTM scores than their working counterparts (p = .029). A positive relationship was found between GPA and CTM, with higher GPAs correlating to higher motivation (p = .002). Students who used ChatGPT showed significantly higher CTM scores than those using other AI applications (p = .003). Interestingly, those who attended AI training courses had lower CTM scores (p = .021). Gender, residence, AI usage duration, and sources of information about AI did not reveal significant differences in CTM scores (Table 4).

Table 4.

The Relationship Between Students’ Socioeconomic Characteristics and CTM.

Variables		n	Mean ± SD	t/f	p-value
Age			20.49 ± 1.20	.044	0.384
Academic year
	First	22	71.72 ± 21.66	5.624	<.001*
	Second	143	85.19 ± 19.28
	Third	147	87.63 ± 18.26
	Fourth	78	89.06 ± 15.63
Gender
	Male	290	87.16 ± 19.03	1.872	0.052
	Female	100	83.12 ± 17.40
Residence
	Urban	322	87.14 ± 18.26	3.008	0.051
	Semiurban	36	82.72 ± 22.15
	Rural	32	79.68 ± 17.59
GPA			4.30 ± .58	.159	0.002*
Do you work during your study?
	Yes	49	80.69 ± 21.93	−2.187	0.029*
	No	341	86.90 ± 18.08
What are your sources of information about AI?
	Internet	305	86.39 ± 19.19	.312	0.732
	Friends	71	84.64 ± 17.15
	Workshop	14	87.85 ± 15.50
What are these applications?
	ChatGPT	217	87.85 ± 17.82	5.973	0.003*
	Deepseek	95	80.44 ± 22.26
	Gmeini	78	88.25 ± 14.71
Hours use
	1	223	87.03 ± 16.54	.495	0.739
	2	74	83.75 ± 23.81
	3	58	86.58 ± 16.87
	4	24	84.75 ± 23.14
	5	11	84.27 ± 21.23
Since when have you been using AI applications?
			1.41 ± .76	.015	0.766
Have you attended training courses about AI usage in nursing previously?
	Yes	66	81.28 ± 19.08	−2.321	0.021*
	No	324	87.11 ± 18.48

Note. CTM = critical thinking motivation; t = Student t test; f = one-way analysis of variance test; GPA = grade point average; AI = artificial intelligence. *Statistically significant at p ≤ .05.

Multiple linear regression revealed a significant model (p < .001) when sociodemographic, AI attitudes, AI usefulness, and AI usability were analyzed as predictors of CTM. The variance of 27.0 in CTM was explained by the model (R² = .270, adjusted R² = .245). Being in the second, third, and fourth levels was a significant predictor of CTM when compared to the reference category. In addition, students who did not think that AI may replace nursing staff in the future were a significant factor with higher CTM than the reference category. Finally, AI usability was a predictor of higher CTM (Table 5).

Table 5.

Factors Affecting Nursing Students’ Critical Thinking Motivation.

Factors	Categories	Unstandardized beta (β)	SE	Standardized beta (β)	t	p	95% CI for β
Academic level	First	Reference
	Second	13.70	3.76	.35	3.63	<.001*	6.29-21.10
	Third	14.11	3.81	.36	3.70	<.001*	6.61–21.60
	Fourth	16.40	4.04	.35	4.05	<.001*	8.45–24.35
GPA		2.33	1.47	.07	1.58	.114	−.56 to 5.24
Working during study	Yes	Reference
	No	1.59	2.65	.02	.60	.549	−3.62 to 6.80
Application use	ChatGPT	Reference
	Deepseek	−3.46	2.10	−.08	−1.64	.100	−7.60 to .66
	Gemini	1.59	2.23	.03	.71	.477	−2.80 to 5.99
Courses and training	Yes	Reference
	No	−1.17	2.44	−.02	−.48	.632	−5.96 to 3.62
Thinking to incorporate AI in the study	Yes	Reference
	No	−2.03	1.92	−.05	−1.06	.289	−5.81 to 1.73
AI may replace nursing staff in the future?	Yes	Reference
	No	10.14	2.27	.22	4.45	<.001*	5.66–14.61
AI attitudes		−.19	.13	−.08	−1.43	.152	−.45 to .07
AI usefulness		.18	.29	.04	.61	.541	−.39 to .75
AI usability		1.46	.31	.34	4.60	<.001*	.84–2.09

Note. GPA = grade point average; AI = artificial intelligence; R²= .270; adjusted R²= .245; p < .001; CI = confidence interval. *Significant level < .05.

Discussion

The rapid integration of AI into education has prompted a critical inquiry: Does AI enhance nursing students’ motivation for critical thinking? As AI tools such as ChatGPT become increasingly prevalent in academic settings, understanding their impact on higher-order cognitive skills is essential. This study, conducted through a mixed-methods approach involving surveys and interviews, examines the relationship between AI perceptions—attitudes, usefulness, and usability—and CTM among Saudi nursing students.

The primary finding of this study was the generally high positive attitude toward the use of AI among Saudi nursing students. The theory of reasoned action posits that behaviors are determined by an individual's predisposition toward a specific action, influenced by personal attitudes and subjective norms (Fishbein & Ajzen, 1977). This favorable perception can be attributed to several factors. Over recent years, Saudi Arabia has made substantial progress in AI development, closely aligned with the goals of Vision 2030, which emphasizes digital transformation across various sectors, including healthcare and education (Vision 2030, 2023). The integration of AI into academic and clinical environments has enhanced students’ exposure to these technologies, promoting familiarity and acceptance.

Furthermore, the widespread availability of open-access digital resources, a robust technological infrastructure, and high-speed internet access across the Kingdom have likely contributed to students’ engagement with AI applications (Alruwail et al., 2025; Mihmas Mesfer Aldawsari & Rashed Ibrahim Almohish, 2024).

These findings are consistent with the existing literature. For example, Migdadi et al. (2024) reported moderate levels of ethical awareness, attitudes, anxiety, and intention to use AI among Jordanian nursing students, reflecting a growing regional acceptance of AI technologies in nursing education. Similarly, a multicenter cross-sectional study by Al Omari et al. (2020) involving a convenience sample of 1,713 nursing students found that most participants demonstrated moderate levels of knowledge, attitudes, perceptions, and intentions regarding the use of AI. Notably, there was a significant positive association between students’ knowledge, attitudes, perceptions, and their intention to utilize AI in clinical practice.

While our findings demonstrate statistically significant but weak correlations between AI attitudes/usefulness/usability and CTM, such modest effect sizes suggest that positive perceptions of AI alone may not substantially enhance critical thinking engagement. This aligns with broader literature cautioning that overreliance on AI tools can lead to cognitive offloading, reducing mental effort, and undermining independent analytical processes (Zhai et al., 2024). For instance, an MIT EEG study revealed that frequent ChatGPT users showed lower neural engagement and creativity than those writing unaided by AI (Time, 2025). These findings suggest that while students appreciate AI's convenience, its use may circumvent the deeper cognitive struggle required for critical thinking.

These AI-driven scenarios replicate real-world clinical situations and allow students to repeatedly practice decision-making and problem-solving in a safe learning environment—a key advantage of virtual simulations in healthcare education (Alammari, 2024). Complementing these findings, Gerlich (2025) found that while frequent AI use correlates with reduced cognitive effort, this may inadvertently undermine critical thinking performance due to increased reliance on automated processing rather than engaging in independent reasoning (Gerlich, 2025).

The regression analysis indicated that higher academic levels, the perception that AI does not pose a threat to or replace nursing roles, and an increased perceived usability of AI were significant predictors of enhanced motivation for critical thinking. These findings can be effectively interpreted through the framework of Davis (1989). The TAM suggests that perceived usefulness and perceived ease of use are primary drivers of technology acceptance and behavioral intention. In this context, the perceived usability of AI reflects ease of use. At the same time, the belief that AI supports rather than replaces nursing functions aligns with perceived usefulness, both contributing positively to users’ CTM.

Regarding the counterintuitive finding that attendance in AI training courses was associated with lower CTM, several mechanisms may elucidate this pattern. First, students emerging from structured training may experience cognitive dissonance; they acquire knowledge about AI capabilities and limitations, which could temporarily heighten their awareness of AI's shortcomings and undermine confidence, thus decreasing motivation (Habib et al., 2025). Second, AI training may accentuate automation bias, wherein trained individuals become aware of AI's fallibility but lack the expertise to critically evaluate outputs, leading to both reliance and skepticism that diminish engagement (Fan et al., 2025). This effect is supported by studies indicating that access to AI can paradoxically reduce motivation for independent effort, described as “metacognitive laziness” (Fan et al., 2025; Habib et al., 2025).

This interpretation aligns with the findings of Jia and Tu (2024), who reported that AI capabilities can indirectly foster critical thinking by enhancing students’ self-efficacy and learning motivation. Their study emphasized that self-efficacy plays a pivotal mediating role, influencing both learning motivation and critical thinking awareness. These insights suggest that although AI has the potential to reshape cognitive engagement, its impact may be mediated by users’ perceptions and beliefs, as described in TAM. This reinforces the idea that positive attitudes and confidence toward AI tools are essential to fully leveraging their educational benefits.

Cross-cultural comparisons further illuminate the interplay between AI attitudes and CTM. For instance, a study by Jiang (2024) in China revealed that nursing students expressed high acceptance of AI as a learning enhancer, particularly when integrated through gamified, simulation-based platforms. This high acceptance was partially attributed to China's national digital literacy campaigns and students’ early exposure to AI in secondary education.

Moreover, a major UK survey by the Higher Education Policy Institute (2025) reported that 92% of undergraduate students now regularly use generative AI tools such as ChatGPT, up from 66% the previous year. While students value AI's utility for time saving and clarifying academic ideas, UK academics have raised concerns that this growing reliance may diminish students’ capacities for independent critical thinking, especially among undergraduates whose responses tended to improve only cosmetically when tasked with enhancing AI-generated content. These findings mirror our results, underlining the importance of fostering strong foundational reasoning skills so that students can use AI as an aid rather than a crutch. In conclusion, our study highlights the need for a balanced approach to AI integration in nursing education, where AI is seen as a tool to enhance learning and critical thinking, rather than a replacement for independent reasoning.

In the United States, Pitts et al. (2025) investigated the utilization of AI chatbots at a large public university, revealing that while students valued the immediate feedback and access to information, they harbored concerns regarding academic integrity, information accuracy, overreliance on AI, and the potential erosion of critical thinking skills—concerns that resonate with those observed in our Saudi cohort. Conversely, a recent Brazilian study by Nicolás and Sampaio (2024) reported a lower overall acceptance of AI, attributed to infrastructural limitations and apprehensions about data privacy, thereby highlighting the influence of national readiness and digital infrastructure on students’ attitudes.

These international insights corroborate the central conclusion of the current study: that student attitudes, perceived AI usability, and the cultural framing of AI's role in healthcare significantly influence motivation for critical thinking. They further emphasize the necessity of context-specific curriculum strategies that account for local technological exposure, cultural values, and educational priorities when integrating AI into nursing education.

Strengths and Limitations

This study offers several strengths, including its multicenter design, which involved three universities in Saudi Arabia, allowing for a diverse sample of nursing students from different regions and institutions. The use of well-established and validated measurement tools such as the GAAIS, TAM, and CTMS enhances the reliability and robustness of the findings. With a large sample size of 664 participants, the study ensures strong statistical power and generalizability to the broader population of nursing students in Saudi Arabia. Additionally, the study adheres to ethical standards with appropriate informed consent and data security protocols, ensuring participant confidentiality.

However, the study also has limitations. The cross-sectional design, while useful for capturing attitudes and motivations at one point in time, prevents establishing causal relationships. The reliance on self-report measures introduces the possibility of biases such as social desirability or recall bias. The convenience sampling method, though practical, limits the generalizability of the findings, as the sample may not fully represent the broader nursing student population. Moreover, the study was conducted only across three universities, which may not fully capture regional or institutional variations. Lastly, the lack of in-depth exploration into students’ actual AI usage and other potential influencing factors limits the scope of the findings. Despite these limitations, the study offers valuable insights into the role of AI in nursing education and its potential to enhance CTM.

Nursing Implications

Nursing educators, academic institutions, and policymakers must take intentional steps to translate positive attitudes toward AI into enhanced cognitive outcomes. The findings of this study offer the following concrete implications: Early exposure to AI tools (e.g., ChatGPT, DeepSeek, and virtual simulation platforms) in the foundational years of nursing education, particularly in years 1–2, has been shown to significantly boost CTM in later academic stages. Therefore, it is recommended that AI applications be gradually integrated into clinical courses. Assignments should be redesigned to necessitate student interaction with AI tools, not for direct answers but for generating multiple clinical perspectives, differential diagnoses, or ethical considerations. For instance, students could compare AI-generated care plans with evidence-based guidelines and critically evaluate any discrepancies.

Institutions should provide formal workshops on how to use AI ethically, accurately, and effectively. These should include bias awareness, citation standards, prompt engineering, and recognizing hallucinations, particularly for faculty who may lack confidence in integrating AI into their teaching. Our data show that students who do not believe AI will replace nurses have higher CTM. Therefore, educators play a crucial role in shaping students’ attitudes toward AI, emphasizing their role as a clinical reasoning assistant, not as a decision-maker. This should be reinforced in both lectures and assessments.

It is recommended to create blended activities where students use AI to generate case variables or progress notes, followed by engaging in peer discussion and critique. This approach fosters collaborative problem-solving and prevents passive reliance on AI output. Institutions should establish clear academic policies for acceptable AI use, including transparency in using AI for assignments, the requirement to disclose AI assistance, and training students on academic integrity in the age of generative tools. Collaboration with computer science or health informatics departments to codevelop nursing-specific AI tools is also encouraged, as it aligns digital tools with clinical relevance and increases usability perceptions—an important predictor of CTM, as per our findings. Students should be encouraged to maintain AI usage logs or reflective journals documenting how they used AI, what they learned, and where they encountered limitations. This practice promotes metacognitive awareness and reinforces self-regulated learning.

By implementing these strategies, nursing educators can maximize AI's pedagogical potential while fostering a culture of analytical inquiry, ethical judgment, and technological fluency—hallmarks of a future-ready nursing workforce.

Conclusions

The study revealed that nursing students primarily used ChatGPT, followed by DeepSeek and Gemini, in their academic activities. Students generally held positive views toward AI, as shown by their moderate to high scores in AI attitudes, usefulness, and usability. CTM was in the moderate to high range, with fourth-year students scoring significantly higher than their peers. Strong correlations were found among AI attitudes, usefulness, and usability, while weak to moderate correlations were observed between these AI-related factors and CTM. Educators and curriculum developers should prioritize strategies that improve students’ perceptions of AI's usefulness and ease of use to maximize its educational benefits.