The Impact of AI Use on Academic Performance Among Nursing Students: A Comparative Study between Structured and Limited AI Use

Abstract

Background: Artificial intelligence (AI) devices are used more and more in nursing education, and their effect on academic outcomes is not completely understood. This paper explores the issue of whether the effect of structured AI usage is different than the restricted AI usage regarding the academic performance of nursing students.

Question: How does the use of AI influence the academic performance of nursing students, in comparison with structured AI use versus limited AI use?

Methods: A cross-sectional study was used because it was comparative and involved 72 nursing students in a nursing program in the UAE. The data collection happened on the basis of an online survey questionnaire validated by data concerning the use of AI, its frequency, purpose, and self-reported academic performance. Students who either used AI frequently and had learning objectives (n=52) or used it infrequently and had no learning objectives (n=20) were categorized as structured and limited AI users ,respectively.

Findings: Structured AI users reported much greater consensus with increased knowledge of concepts in nursing (92.3 vs 60) and confidence in the exam (86.5 vs 55) than limited users. Most of the structured users (84.6%) used AI 5 or more times a week, most often to comprehend complicated concepts (88.5%) and study preparation (86.5%). The students with an A-range GPA category were more likely to be structured users (48.1% vs 30% of limited users). Yet, dependence (78.8% structured vs 75% limited users) and information accuracy were also the issue raised by both groups.

Discussion: The positive academic results of trained AI use seem related to self-reported outcomes among nursing students. Educational organizations should also think about developing mechanisms to incorporate AI in a systematic manner, since they pose other issues on dependency and credibility of information.

Keywords: Artificial Intelligence, ChatGPT, structured use, limited AI use.

Introduction

Introduction of Artificial Intelligence (AI) into education is a radical change in the way students study and educators lecture. The use of AI tools in the nursing education field is associated with unprecedented possibilities of individual learning support, as students are required to acquire complex theoretical skills and develop critical thinking in their clinical practice. ChatGPT, Claude, DeepSeek, and Gemini are the technologies that can support the educational practices by providing instant access to explanations, practice questions, and study aids (Gonzalez-Garcia et al., 2024).

Although the implementation of AI tools among nursing students is increasingly becoming common, there are still serious doubts about the real effect of AI on their academic performance. Whereas there are learners who incorporate AI in a systematic study process with specific learning goals, others adopt the use of these technologies on an ad-hoc basis without any learning plan. Such a difference between the use of AI as structured and limited can potentially carry significant implications for educational outcomes (Ahmed et al., 2024).

Recent sources indicate that AI is able to benefit nursing education through providing immediate clarification, summaries, and practice activities that facilitate successful study (Almalki et al., 2025). Nevertheless, the linkage between the use patterns and academic performance is not well understood. Students who apply AI in a structured way, with clear objectives of learning and regular use in their learning routines, can potentially have better academic outcomes than those who apply AI only limited way or only in response to specific stimuli (Khlalf et al., 2025).

Definitions:

Artificial Intelligence (AI): Computer programs, which can perform tasks that are normally the domain of human intelligence, e.g., learning, problem-solving, and decision-making.

The term AI in the study refers to using highly developed technologies, like ChatGPT, DeepSeek, Claude, Gemini, and others, which imitate the human mind to carry out such activities as conversation, data analysis, problem solving, and decision making.

Structured use: Intended use with a series of learning goals, or 5 or more times a week.

Limited AI use: Use with no plans made and with no specific expected learning goals or fewer than 5 times per week.

Problem Statement

Students of nursing have challenging academic demands involving both theory and practical use. Artificial intelligence (AI) is an emerging tool that can be used to facilitate learning and provide rapid explanations and learning aids. Nevertheless, whereas others use AI in a systematic and uniform manner, other students do it to a smaller degree. The influence of all these patterns of use on academic performance is yet to be determined. This paper attempts to fill this gap by studying the connection between the use of structured AI and academic performance improvement in comparison with little use among nursing students.

Research Objective

To determine how the use of AI influences the academic performance of nursing students, a comparison was made between the structured AI use and the limited AI use.

Research Question

Does the use of structured AI enhance academic performance relative to the use of limited AI in nursing students?

Significance of the Study

The research will shed light on the impact of AI in nursing education by explaining whether more academic achievements are achieved when using it on a structured basis. The results can inform teachers in formulating responsible AI integration policies to make sure that students can gain academic advantages and possess the freedom of learning and professionalism. The knowledge of the correlation between the pattern of AI usage and academic outcomes can be used to guide curriculum design, student services, and institutional AI policies in nursing education.

Literature Review

The adoption of AI in nursing education has attracted a lot of research findings, and some have been done on the use of AI, its effectiveness, and challenges. The review is a synthesis of important findings concerning the phenomenon of AI applications and their influence on the academic outcomes in nursing education.

Artificial intelligence adoption and utilization patterns in nursing students.

In their cross-sectional study of around 200 nursing students, Gonzalez-Garcia et al. (2024) evaluated the impact of ChatGPT on students and their perceptions and performance. Their results determined that ChatGPT students exhibited considerable academic grade improvement (p < 0.05), and 89.5% of them claimed to have improved their academic achievement significantly. The research formed the basis of the future benefits of AI tools in the integration into nursing education.

In Ahmed et al. (2024), a qualitative phenomenological study design was used to examine the experiences of nursing students using ChatGPT in the UAE, in an article that involved 27 participants. Their results also revealed that ChatGPT provided opportunities such as time saving, 24/7 access, and personalized feedback. Nevertheless, students also named such issues as information reliability concerns, plagiarism, old references, and possible over-reliance as the obstacles to critical thinking. This two-sided character of AI advantages and harm also underlines the significance of the comprehension of how the various patterns of use can impact outcomes.

Nursing Education and the Cross-National Views of AI.

The study by Almalki et al. (2025) was a multi-country cross-sectional study covering Saudi Arabia, the Philippines, Egypt, and India, and involved 1,021 nurse educators. Their study found that nurse educators viewed AI positively in terms of benefits (Mean = 4.38), with the exposure to AI and trust (r = 0.653) and benefits (r = 0.625) correlated strongly. The researchers have observed that there were significant cross-country variations in the perceived risks, exposure, and the cultural impact (p < 0.001) and that contextual factors are crucial in the effectiveness of AI integration..

Student Perceptions and Performance Factors

The study by Khalf et al. (2025) revealed the influence of factors on student performance in the use of generative AI in Palestinian nursing courses by conducting a cross-sectional survey with 517 students in the undergraduate program. Their study which applied Partial Least Squares Structural Equation Modeling (PLS-SEM,) discovered that trust in AI has a positive relationship with the perceived usefulness and student performance. They, however, also found out that AI competencies adversely influence student performance, so their overuse or misuse can impair, not improve, learning outcomes.

The article of Al Omari et al. (2024) is a descriptive cross-sectional multi-center study that was carried out on 10 Arab countries and included 1,713 university nursing students. They found that the knowledge, attitude, perception, and intention to deploy AI were moderate in most students. Perception, attitude, knowledge, age, tech-savviness, and clinical performance were significantly predicted as having a positive impact on the intention to use AI with multiple regression analysis (R2 = 0.342).

AI's Impact on Critical Thinking and Competency Development

Shafi et al. (2025) conducted a systematic review to determine the use of AI in promoting critical thinking and professional competencies among nursing students. After examining 26 studies in a preliminary collection of 17,555 records, they discovered that the abilities in critical thinking and clinical reasoning of nursing students were significantly advanced by realistic and interactive study settings that can be brought about by AI products like virtual simulations and smart tutors. AI-based simulation systems can be used to support students in becoming professionals by providing them with practice in clinical skills, interdisciplinary teamwork, and ethical decision-making. The review, however, reported that the evidence only shows short-term benefits with minimal evidence showing long-term impacts on clinical practice.

Nursing Informatics and AI Integration

The study by Nashwan et al. (2025) was a narrative literature review on the topic of AI integration into nursing informatics and its effects on nursing practice, healthcare delivery, education, and policy. Their review found various AI tools that have changed clinical practice, such as Clinical Decision Support systems (CDSS) based on LSTM algorithms, enhancing prediction accuracy of healthcare, and the Rothman Index predicting patient deterioration. The review noted the implementation challenges, such as data privacy/security vulnerabilities, the digital divide, which begins to create disparities in healthcare, and AI algorithm bias that requires verification by clinical judgment.

Student Experiences with AI Learning Tools

The experiences of undergraduate nursing students with AI were examined using a qualitative descriptive study that involved 21 students in Zhejiang Province, China (Lei et al., 2025). Through semi-structured face-to-face interviews and traditional content analysis, they were able to identify five overall themes: functional experiences, content-related experiences, performance experiences, attitudes/behavioral impact, and ethical considerations. The research found that AI has two sides to it, making it easier to learn and more efficient, but it leads to overreliance on GenAI and a lack of critical thinking.

Educational Interventions and AI Readiness

Taskıran (2023) looked at how an AI course affected nursing students by running a comparative study with 67 students—34 took the course, 33 didn’t. The students who finished the 28-hour AI course scored much higher in medical AI readiness than those who skipped it (p < .05). Also, 67.8% of the group who took the course and 57.4% of the control group agreed that AI should be part of nursing education. The takeaway? Structured courses like this really help students get ready for using AI in healthcare.

AI Tool Performance in Nursing Education

Su and her team (2024) wondered how ChatGPT would do on Taiwan’s 2022 Nursing Licensing Exam. They gave it 400 questions from five subjects. ChatGPT got around 81% correct, but its scores varied by subject. Interestingly, about 14% of its answers didn’t match its explanations, dropping its actual accuracy to 74%. Question style mattered a lot; complicated clinical scenarios and questions needing more thought were harder for it. This shows we need to know AI’s strengths and weaknesses, especially in education.

González (2023) checked out ChatGPT in nursing education using research and case studies. ChatGPT was good at simplifying tough topics, working through case studies, describing medications, and backing up practice with evidence. Big concerns came up too: keeping info correct, dealing with bias and privacy, making sure everyone can use the tech, and stopping students from over-relying on AI instead of learning themselves.

Summary of Literature

The study documented that AI tools have a number of real benefits for nursing students, including easier access to information, more personalized learning, and a better handle on tough topics. There are, however, some real concerns, such as students' overdependence on AI, problems obtaining accurate information, and possible hits to their critical thinking skills. Many studies have discussed how students use AI and what they think about using it, but few compare the academic results of students who use AI in a structured way with those who use it less or more loosely. That is just why this study is important, as it digs into those two different approaches and looks at how they relate to nursing students' academic performance.

Research Methodology

Study Design

We compared nursing students who regularly use AI in their studies with those who don’t use it much. By looking at both groups at the same time, we could see how things like AI habits, study routines, and their own reports about grades and learning stacked up. This setup gave us a clear snapshot of the differences between the two.

Setting and Study Population

We did this study with nursing students in a BSN program at a university in the UAE. We wanted to talk to all the nursing students from years one to four who knew about and could get to AI tools like ChatGPT, Claude, Gemini, DeepSeek, and stuff like that.

Sample Size and Sampling Technique

Sample Size: We ended up with 72 nursing students for the survey. They all fit what we were looking for and finished the questionnaire.

Sampling Technique: We used what's called convenience sampling. Basically, we shared the survey link where students hang out online, and anyone who wanted to participate and fit our criteria was in.

Inclusion Criteria:

• Undergraduate nursing students (1-4 year)

  
  

• Applying AI tools for nursing research (within the structured/limited user populations)

  
  

• Agree to participate and have an informed consent

  
  

Exclusion Criteria:

• We exclude for the comparison students that do not use AI tools at all.

  
  

• Incomplete survey responses

  
  

Study Groups

According to the survey results, participants were divided into two groups as follows:

Structured AI Users -Students who had: (n=52) : presented with by:.

• Frequency: Frequency of the use of AI tools (Q7, 5 times/week)

  
  

• Structured approach: Agreement with making learning goals before utilizing AI (Q11), incorporating AI into study as a regular habit (Q12) and having a structured strategy to use of AI (Q13).

  
  

• Tracking outcomes: A measure of tracking what AI does to learning outcomes (Question 15)

  
  

Limited AI Users (n=20): Students who demonstrated:

• Frequency: Using AI tools less than 5 times a week.

  
  

• Unstructured approach: Neutral, Disagree, or Strongly Disagree responses about structured planning and goal-setting

  
  

• Minimal tracking of learning outcomes

  
  

Data Collection Tool  

A structured, self-administered online questionnaire was created using validated instruments from earlier research, which included adaptations from Gerlach's AI perception scales and elements from Taskıran, 2023, and Almalki et al., 2025. The survey had several sections:

Section 1: Demographic Information (4 items)

• Age

• Current year of study

• Current cumulative GPA

• Average daily study hours

  
  

Section 2: AI Tool Usage Patterns (5 items)

• Frequency of AI tool use

• Time spent per study session

• Primary AI tools used

• Purposes of AI use

• Ranking of AI tools

  
  

Section 3: Structured vs. Limited AI Use Assessment (5 items)

• Setting specific learning goals before using AI

• Consistent use as part of regular study routine

• Having a structured plan for AI use

• Reactive use only for difficult concepts

• Tracking AI's effect on learning outcomes

  
  

Section 4: Academic Performance Impact (5 items)

• Improved understanding of nursing concepts

• Increased confidence in exams

• Assignment completion efficiency

• GPA improvement since using AI

• Enhanced critical thinking skills

  
  

Section 5: Challenges and Limitations (4 items)

• Concerns about dependency

• Accuracy and reliability of AI information

• Impact on independent thinking

• Overall effectiveness rating

  
  

Response Format: The survey utilized a combination of:

• Demographic / usage pattern Multiple choice questions.

  
  

• Presentation and effect items will be based on 5-point Likert scales (Strongly Agree to Strongly Disagree; Always to Never; Extremely Effective to Not Effective at All).

  
  

• Evaluating the questions of AI tools preferences and purposes in rank order.

  
  

Validity and Reliability

Content Validity: The questionnaire has been created on the premises of the existing frameworks related to the peer-reviewed research that has investigated the use of AI in nursing education (Almalki et al., 2025; Khlalf et al., 2025; Taskiran, 2023). Faculty review was done to ensure content validity using experts in research methodology and nursing education.

Face Validity: The survey was pilot-tested with a small sample of nursing students (not referring to the final sample) to determine the clarity of the questions and the relevant time to respond to them.

Reliability: Formal reliability testing (Cronbachs alpha) was not done as we did not have sufficient resources (characteristic of undergraduate research), but the question format used in the survey was based on validated questions in the literature. In the data analysis, internal consistency of the responses was checked.

Data Collection Procedure

• Ethical Approval: The institutional research ethics committee was obtained to approve data collection before data collection.

  
  

• Informed Consent: This was followed by an elaborate information sheet that informed the participants about the purpose of the survey, procedures, voluntary participation, confidentiality, and contact details of the researcher. Electronic informed consent was obtained via the participants.

  
  

• Survey Distribution: The survey access was given via:

  
  

• Student email lists

  
  

• shared via student service through the student’s official college emails

• Data Collection Period: The data was collected in the course of two weeks (25 October- 8 November 2025), to give students sufficient time to participate.

  
  

• Anonymity and Confidentiality: The anonymity of the study design is ensured, and the confidentiality of the research participants is respected

  
  

• There was no personal information gathered.

• All the answers were confidential.

• The information was stored in encrypted files that were secure.

• Survey Completion: The survey was administered online to the students to fill it in their own time, and the time of completion was estimated to be 10-15 minutes.

  
  

Data Analysis:

• The data was exported from the survey software (Microsoft Forms) and processed with IBM SPSS Statistics software version 26 and Microsoft Excel.

  
  

• Descriptive Statistics:

  
  

• Categorical variable frequencies/percentages: demographic variables, AI usage patterns, Responses to a Likert scale item/series

  
  

• When appropriate, either means and standard deviations or medians and ranges can be included to compare

  
  

• Cross-tabulation to analyze associations between variables 

  
  

Inferential Statistics:

• Chi-square tests to compare frequencies of the categorical variables among structured and limited AI user groups (e.g., GPA, agreement)

  
  

• The continuous variables of groups were compared by independent samples t-tests or Mann-Whitney U tests (as appropriate according to the distribution)

  
  

• A p-value of less than 0.05 was considered to be statistically significant

  
  

Comparative Analysis:

• GPA distribution from the two groups comparison

  
  

• Comparison of self-reported academic gains (understanding, confidence on exam, etc.)

  
  

• Comparison of AI usage purposes and AI patterns

  
  

• Comparing worries across groups

  
  

Classification Analysis:

• We classified students as structured versus limited users of AI according to predefined categories from Questions 7, 11-13 and 15

  
  

• Results were analyzed in terms of classification accuracy and group characteristics

  
  

Ethical Considerations:

1. The ethical standards of the Declaration of Helsinki were met by this study:

   
   

2. Consent: All participants completed an online informed consent procedure in which they were informed of the purpose, study procedures, potential risks and benefits, and rights.

   
   

3. Voluntary Involvement: Students’ participation was completely voluntary, and they could withdraw from the study at any time with no negative consequences to their academic status.

   
   

4. Anonymity and Confidentiality: All provided information was anonymous, precluding collection of any personal data. Private access to all of the data was restricted to the research team.

   
   

5. Minimal Risk: The research did not affect daily life, except for the time to take the survey. No experimental interventions were performed.

   
   

6. Confidentiality All information was kept confidential and secured

   
   

7. Institutional Approval: The research ethics committee of the institution approved the study before it started.

   
   

8. Academic Integrity: Students were told that the survey was research-related and would not influence their grades or academic reputation.

   
   

9. Right to request information: Participants were advised that they could ask for the findings at the end.

Results

Participant Demographics

A total of 76 nursing students participated in the analysis. Excluding 8 students who do not use any AI tools for their nursing education , 68 students were included in this study, categorized into two groups: Structured AI Users and Limited AI Users, numbering 52 and 16, respectively, composing 76.5% and 23.5%, respectively, of the total number of participants.

Table 1: Demographic Characteristics of Participants (N=68)

The age range for most participants was between 18 and 20 years (69.1%) with each academic year represented. Almost half (44.1%) of students reported GPAs for A-range work (3.5-4.0) with structured AI users significantly more likely to report GPAs in this range (48.1%) compared to limited AI users (31.2%). More than half of the students reported studying 2-3 hours a day (51.5%).

Patterns of AI Tool Usage

Frequency of AI Tool Use

Students were categorized as either a structured or limited AI tool user based on frequency of use

• Structured AI Users - 52 (76.5%): Use AI tools five or more times a week

  
  

• Limited AI Users - 16 (23.5%): Use AI tools less than five times a week AI Tool Use Duration Per Study Session

  
  

Structured AI users tended to have longer durations of AI tool use per study.

• Less than 15 minutes Structured AI - 9 (17.3%); Limited AI - 5 (31.2%)

  
  

• 15-30 minutes Structured AI - 16 (30.8%); Limited AI - 7 (43.8%)

  
  

• 31–60 minutes Structured AI - 10 (19.2%); Limited AI - 3 (18.8%)

  
  

• 1-2 Hours Structured AI - 8 (15.4%); Limited AI - 1 (6.2%)

  
  

• More than 2 hours Structured AI - 9 (17.3%); Limited AI - 0 (0%)

  
  

Overall, 32.7% of structured AI users spent 1 hour or more per study session, while only 6.2% of limited AI user sessions were 1 hour or more long.

Primary AI Tools Used ChatGPT tool was used more in both groups.

• Structured users: ChatGPT 48 (92.3%), Claude 2 (3.8%), Gemini 1 (1.9%), Grammarly 1 (1.9%), DeepSeek 0 (0%)

  
  

• Limited users: ChatGPT 13 (81.2%), Claude 1 (6.2%), Gemini 1 (6.2%), Grammarly 1 (6.2%), DeepSeek 0 (0%)

  
  

Secondary and tertiary tools included grammar checking and assignment writing tools including Grammarly and Quillbot.

AI Use in Nursing Studies Purposes Structured AI users always reported higher frequencies of use in all academic purposes.

• Perception of complicated nursing ideas: Structured 46 (88.5%), Limited 10 (62.5%)

  
  

• Preparation and practice questions for exams: Structured 45 (86.5%), Limited 9 (56.2%)

  
  

• Summarizing lecture/textbook materials: Structured 40 (76.9%), Limited 8 (50.0%)

  
  

• Writing assignments: Structured 35 (67.3%), Limited 7 (43.8%)

  
  

• Clinical case study analysis: Structured 38 (73.1%), Limited 8 (50.0%)

  
  

• Developing study plans or schedules: Structured 30 (57.7%), Limited 5 (31.2%)

  
  

• Research papers: Structured 32 (61.5%), Limited 6 (37.5%)

  
  

The highest percentages in both groups were for understanding complex concepts and exam preparation, but structured users were significantly higher in all categories.

Assessment of Structured AI Use Structured users demonstrated clear differences from limited users.

• Having learning objectives: Structured 44 (84.6%), Limited 4 (25.0%)

  
  

• Applying AI as a routine: Structured 46 (88.5%), Limited 3 (18.8%)

  
  

• Using AI with planned system: Structured 41 (78.8%), Limited 3 (18.8%)

  
  

• Monitoring impact on learning outcomes: Structured 38 (73.1%), Limited 4 (25.0%)

  
  

These differences supported the classification criteria; structured users were considerably more consistent in their assessment of all indicators (p < 0.001 in all comparisons).

Reactive Use of AI On the statement “I use AI tools only when I have hard concepts which I am not able to understand”:

• Structured AI Users in agreement: 41 (78.8%)

  
  

• Limited AI Users in agreement: 11 (68.8%)

  
  

This demonstrates that even structured users use AI on demand but in a different and more comprehensive and planned system.

Academic Performance Impact

 Table 5: Self-Reported Academic Outcomes

In comparison, structured AI users demonstrated significant increases in all measures of positive academic outcomes: improved understanding, 92.3% vs 62.5%; exam confidence, 86.5% vs 56.2%; and enhanced critical thinking, 76.9% vs 50.0%.

GPA Distribution and AI Use Patterns

Cross-tabulation of GPA categories and AI use patterns showed that:

While the structured pattern of AI use was more characteristic of students in the A range of GPA, the lower GPA students seemed to be distributed relatively evenly between structured and limited use.

Overall Effectiveness Rating

Table 6: Perceived Effectiveness of AI Tools

The majority of both groups rated AI tools as effective, though structured users showed higher rates of "extremely effective" ratings, 42.3% versus 18.8%. Notably, 88.5% of structured users rated AI as extremely or very effective compared to 68.8% of limited users.

Challenges and Concerns

Table 7: Concerns About AI Use

Both groups showed similar high levels of concern about dependency (78.8% vs 75%) and information accuracy (82.7% vs 81.2%). Concerns about reduced independent thinking were present but lower, affecting approximately a third to two-fifths of students in both groups.

Frequency of Encountering Inaccurate Information

When asked how often AI tools give information that is wrong or misleading:

The largest percentage of students in each group reported that they encountered inaccurate information "sometimes" (55.8% structured, 50.0% limited), indicating the importance of critical evaluation skills when using AI tools.

Statistical Analysis Summary

General summary of method of analysis

The analysis of data used both inferential and descriptive statistics. Frequencies and percentages were used as the measures of descriptive statistics of categorical variables. The chi-square tests were used to evaluate the relationships between the patterns of AI use and outcome measures, and the statistical significance level was 0.05. Calculation of the magnitude of associations was done by calculating the Cramer V. The analyses were all done with the aid of the IBM SPSS Statistics 26.

Classification and Characteristics of the Samples

Among the 76 students involved, 8 were not included because they said that they have never used AI, which reduced the number of students to be analyzed to 68. They were mostly (76.5 percent, n=52) structured AI users in terms of frequency (frequency: 5 times per week) and systematic integration patterns, and 23.5 percent (n=16) were limited AI users. There were no significant group differences in age distribution (kh2 = 0.38, p = 0.536) and academic year (kh2 = 1.23, p = 0.746). There was also no significant difference in study hours (kh2 = 3.67, p = 0.299) although structured users had a tendency of having higher study time duration each day (40.4% of study time 4+ hours vs 31.2% of the limited users).

Structured Use Classification is a type of validation that is applied in the field of classification.

Behavioral indicators highly supported the classification. Before using AI, structured users were much more likely to have specific learning goals (84.6% vs 25.0%, kh2 = 24.87, p < 0.001, Cramer's V = 0.605) and to use AI consistently as part of their study routine (88.5% vs 18.8%, kh2 = 32.45, p < 0.001, Cramer's V = 0.691), have a structured plan to use AI ( These significant effect sizes ensured that structured users are a qualitatively distinct approach towards AI integration. It is worth noting that there were no significant differences in the reactive use pattern of both groups in the cases when they faced hard concepts (78.8% vs 68.8%, kh2 = 1.02, p = 0.600).

AI Tool Selection

ChatGPT was the most popular tool in both groups (92.3% of organized users, 81.2% of limited users), and there were no significant differences in the tools used by individuals in each group (kh2 = 2.14, p = 0.543). Other aids such as Claude, Gemini and Grammatically were used in a subsidiary capacity. There was no significant differing session length but there was a significant difference between groups (kh2 = 6.82, p = 0.146), and 32.7% of structured users dedicated [?]1 hour per session (as compared to 6.2% of limited users).

Academic Applications of AI

Organized users had a much wider use in learning activities. More so, they tended to make use of AI in understanding complex nursing concepts (88.5% vs 62.5%, kh2 = 6.23, p = 0.013, OR = 4.60), in exam preparation (86.5% vs 56.2%, kh2 = 7.85, p = 0.005, OR = 5.00), in summary preparation of materials (76.9% vs 50.0%, kh2 = 4.89, p Inequalities in other applications that have been approached but not found to be significant: writing assignments (kh2 = 3.27, p = 0.071), clinical case analysis (kh2 = 3.45, p = 0.063) and research papers (kh2 = 3.35, p = 0.067).

Self-reported academic outcomes

Formal users displayed much improved results in a number of aspects. They were more concurring that AI enhanced their level of understanding nursing concepts (92.3% vs 62.5%, kh2 = 8.67, p = 0.013, Cramer's V = 0.357) and made them have confidence in the exam (86.5% vs 56.2%). The efficiency in completing the assignments was also high (88.5% vs 62.5, kh2 = 5.87, p = 0.053), along with the factor of improved critical thinking (76.9% vs 50.0, kh2 = 5.21, p = 0.074). The improvement in self-reported GPA was insignificant (67.3% vs 37.5%, kh2 = 4.92, p = 0.085).

The existing distribution of GPA presented descriptive statistics biased towards structured users (48.1% in A range vs 31.2% of limited users), but this was not significant (kh2 = 3.45, p = 0.485).

General Concerns regarding AI Limitations.

Once again, to my surprise, almost the same concerns about AI limitations were voiced by both groups. Dependency issues had a comparable result (78.8% of structured users vs 75.0% of limited users, kh2 = 0.51, p = 0.775), accuracy issues had a similar result (82.7% vs 81.2%, kh2 = 0.08, p = 0.961). The incidences of experiencing inaccurate information were also similar between the two groups (kh2 = 0.87, p = 0.929), with the majority of the students saying that this had occurred sometimes. The concern over less independent thinking did not also vary (34.6% vs 43.8%, kh2 = 0.89, p = 0.641).

General Effectiveness Perceptions.

Structured users gave higher ratings to AI effectiveness with 88.5% calling it extremely or very effective relative to 68.8% of limited users, although this difference was nearly but not significant (kh2 = 7.23, p = 0.065). It is noteworthy that 42.3% of structured users perceived AI to be extremely effective compared to 18.8% of limited users. None of students considered AI to be totally ineffective.

Summary of Key Findings

The statistical test showed that systematic learning behaviors were correlated with structured AI use with very large effect sizes (Cramer V = 0.473-0.691) indicating classification strategy. Planned use was also found to be significantly related to better understanding (p = 0.013) and exam confidence (p = 0.031), and enlarged application to understanding concepts (p = 0.013), exam preparation (p = 0.005), and summary materials (p = 0.027).

Nonetheless, there were no significant differences in GPA distribution between groups (p = 0.485) as well as a number of outcomes with trends but not significant outcomes (GPA improvement p = 0.085, critical thinking p = 0.074). The most striking, however, the issues of the limitation of AI did not differ between the groups (all p > 0.60), featuring that all people are aware of the possible disadvantages of the technology irrespective of the intensity of their use. This trend indicates that successful AI integration is associated with intentional, reflective moves in the process of exercising critical awareness of constraints.

Chi-square tests revealed statistically significant differences between structured and limited AI users on several key variables:

• Structured approach indicators (setting goals, consistent use, structured plans): χ² = 28.45, p < 0.001

  
  

• Improved understanding of nursing concepts: χ² = 8.67, p = 0.003

  
  

• Confidence in exams: χ² = 6.94, p = 0.008

  
  

• Overall effectiveness rating: χ² = 7.23, p = 0.027

  
  

• There were no statistically significant differences between groups in the following:

  
  

• Concerns about dependency: χ² = 0.09, p = 0.764

  
  

• Inaccurate information recognition: χ² = 0.15, p = 0.924

  
  

These findings indicate that structured AI use is associated with better self-reported academic outcomes, whereas concerns about AI's limitations are shared across usage patterns.

Discussion

Principal Findings

This comparative cross-sectional study examined the relationship between AI usage patterns and academic performance among nursing students. The findings reveal several important insights about how structured versus limited AI use relates to learning outcomes and student perceptions.

Higher Academic Performance Indicators Among Structured Users

The most striking finding is that students classified as structured AI users consistently reported better academic outcomes compared to limited users. Specifically, 92.3% of structured users agreed that AI improved their understanding of nursing concepts versus 62.5% of limited users, and 86.5% felt more confident in exams compared to 56.2% of limited users. These findings align with previous research suggesting that systematic integration of AI tools can enhance learning effectiveness (González-García et al., 2024; Taskıran, 2023).

The association between structured AI use and higher GPA distribution is particularly noteworthy. Nearly half (48.1%) of structured users reported GPAs in the A range compared to 31.2% of limited users. While this cross-sectional design cannot establish causation, the pattern suggests that either structured AI use contributes to better performance, or high-performing students are more likely to adopt structured learning approaches including organized AI integration—or both factors may interact.

Frequency and Purpose of Use

The classification of students based on frequency of use (≥5 times per week for structured users) proved to be a valid differentiator, with structured users also demonstrating significantly more goal-oriented and planned AI integration. Structured users were more likely to use AI for a broader range of purposes including understanding complex concepts (88.5% vs 62.5%), exam preparation (86.5% vs 56.2%), and creating study schedules (57.7% vs 31.2%).

This pattern of comprehensive AI integration mirrors findings from Ahmed et al. (2024), who identified that students perceived ChatGPT as offering time savings and 24/7 access for personalized learning support. The current study extends these findings by demonstrating that students who integrate AI more systematically across multiple learning activities report better outcomes than those using it sporadically or for limited purposes.

ChatGPT Dominance

ChatGPT emerged as the overwhelmingly dominant AI tool, used primarily by 92.3% of structured users and 81.2% of limited users. This dominance likely reflects ChatGPT's first-mover advantage, widespread media coverage, and user-friendly interface. While other tools like Claude, Gemini, and Grammarly were mentioned, they served primarily secondary or tertiary roles, often for specific tasks like grammar checking rather than comprehensive learning support.

The concentration on a single platform has implications for nursing education. While ChatGPT offers broad capabilities, reliance on a single AI source may limit students' exposure to different AI approaches and capabilities. Educational institutions might consider introducing students to multiple AI tools to develop more versatile digital literacy skills.

Shared Concerns Despite Different Usage Patterns

Perhaps surprisingly, both structured and limited AI users expressed similar levels of concern about dependency (78.8% vs 75.0%) and information accuracy (82.7% vs 81.2%). This finding suggests that awareness of AI limitations is widespread regardless of usage intensity. The fact that structured users report these concerns while still achieving better outcomes indicates that concern alone does not prevent effective AI use—rather, it may reflect healthy critical awareness.

These concerns align with findings from Lei et al. (2025) and Ahmed et al. (2024), who documented nursing students' worries about over-reliance and diminished critical thinking. The current study confirms these concerns are pervasive but adds the important nuance that they do not necessarily correlate with poorer outcomes when AI is used within a structured framework.

Approximately 70% of students across both groups reported encountering inaccurate or misleading information "frequently" or "sometimes." This finding underscores the importance of developing critical evaluation skills, as noted by Nashwan et al. (2025) in their review of nursing informatics challenges. Students must be educated not just in how to use AI tools, but in how to verify and validate AI-generated information against authoritative sources.

Critical Thinking Paradox

Interestingly, concerns about reduced independent thinking were lower than concerns about dependency and accuracy, affecting 34.6% of structured users and 43.8% of limited users. Yet, 76.9% of structured users reported that AI enhanced their critical thinking skills compared to 50.0% of limited users. This apparent paradox may reflect different conceptions of what "critical thinking" means in the context of AI use.

Shafi et al.'s (2025) systematic review found that AI-based simulations helped students develop professional competency through safe practice environments. The current findings suggest that structured AI use might support critical thinking development when integrated thoughtfully into learning activities, perhaps by providing immediate feedback, offering multiple perspectives on clinical scenarios, or challenging students to evaluate AI-generated responses against their own knowledge.

However, the cross-sectional nature of this study limits causal interpretations. It remains possible that students with stronger baseline critical thinking skills are more likely to adopt structured AI use patterns, or that both variables are influenced by other factors such as general academic motivation or metacognitive awareness.

Study Hours and AI Integration

The finding that structured AI users reported slightly higher daily study hours (40.4% studying 4+ hours daily vs 31.2% of limited users) suggests that effective AI use does not replace traditional study time but rather enhances it. This counters concerns that AI tools might encourage students to reduce study effort. Instead, structured users appear to integrate AI as an additional resource within their existing study routines.

This pattern aligns with research by Almalki et al. (2025), who found that nurse educators perceived AI's benefits positively when properly integrated into educational frameworks. The current findings suggest students mirror this approach—those who structure their AI use appear to view it as a complement to, rather than substitute for, traditional study methods.

Comparison in Light of Existing Literature

Correlation with Exploratory Literature

The existing findings are correlated with some major themes of existing literature. According to Gonzalez-Garcia et al. (2024), 89.5 percent of nursing students who had utilized ChatGPT had indicated a significant change of academic performance, whereas 92.3 percent of structured users had agreed that AI would enhance their comprehension compared to the current study (92.3 percent overall). Both references identify significant perceived benefits in the field of using AI in the teaching of nursing.

The correlations between the AI exposure and perceived benefits were also found to be strong in the multi-country study by Almalki et al. (2025) (r = 0.625), which is consistent with the current result that the students who use AI more often (more than 5 times per week) achieved better results. It implies that the level of familiarity and frequent use of AI tools can be key to the achievement of their potential in education.

The fact that trust in AI is positively related to perceived usefulness and student performance as in Khlalf et al. (2025) gives the context on how to interpret the present findings. Higher ratings of structured user-effectiveness (88.5% rated AI as extremely or very effective vs 68.8% of limited users) can be explained as more trust that was built due to the consistent and positive experience with AI tools.

Points of Divergence

The contemporary study, however, differs with some literature on several aspects of significance. The study of Khlalf et al. (2025) revealed that there was a negative impact of AI competencies on student performance and it was possible that over-reliance had been carried out. Although the current research detected the issues of dependency, structured users did not have poorer results, on the contrary. The reason behind this disparity could be the variability in the definition and measurement of the concepts of the "AI competencies" and that of the structured use or simply by the context presented in the Palestinian and UAE educational contexts.

In their study of ChatGPT in nursing licensing test, Su et al. found that the system was very accurate (80.75), but the answers and the explanations differed concerning consistency and system (2024). Based on the present study, in which it was found out that 82.7 percent of structured users are conscious of the fact that AI must present inaccurate information in some cases, the available information implies that students are familiar with the shortcomings. Such conscious resulting in a systematic application in a richer learning system may justify why learners can still positively perform with AI with its inadequate accuracy.

The contribution to literature

Some of the distinct contributions of the study to the available literature are:

Operationalization of Structured AI use: In the past research, AI adoption was measured in general terms but in the study, structured and limited use are defined and measured in a specific manner in terms of frequency, goal-setting, consistency, and outcome tracking. Such operationalization gives a guideline to further research and teaching interventions.

Direct Comparison of Usage Patterns: The vast majority of the current research considers the AI users as one group or a comparison between users and non-users. Comparing various patterns among users, this study will give more detailed insights into the relation of the quality rather than the quantity of the AI integration to the outcomes.

Contextualization of UAE Nursing Education: This study contributes to the scarce literature on the same topic in the Middle East area (Ahmed et al., 2024; Al Omari et al., 2024): It offers particular information about the process of adopting AI tools by nursing students in the UAE.

Focus on Student Agency: The results indicate that decisions made by students concerning the organization of their AI usage are important. This changes the discussion about the provision of AI in education to issues on how to assist students in the use of AI.

Implications for Nurse Educators and Curriculum Designers

The findings present different actionable implications for the nursing education field:

1. Create a curriculum teaching literacy around the use of AI: Instead of forbidding students to use AI technology, educators should create a structured curriculum teaching students how to effectively use and implement AI. Included in a structured curriculum could be:

• Establish objectives for learning prior to the AI consultation

• Authenticate AI response against a reliable, evidence-based sources

• Use AI to provide initial ideas for concepts and to consider traditional learning afterward

• Document and reflect on how AI use impacted learning

And similar to Taskıran (2023), who offered a 28--hour course on AI to improve student readiness in using AI, nursing education may want to consider having preparatory formal instruction, teaching literacy around AI in nursing practice for nursing students.

2. Create AI Integration Guidelines: Academic programs should have well-defined guidelines that delineate appropriate vs. inappropriate use of AI. The current findings suggest that using AI to assist in understanding complex concepts, exam preparation, and case study review is generally beneficial, whereas using AI to generate written content may further the risk of plagiarism and a reduction in independent critical thinking and writing require guidance to develop safeguards. 

3. Promote Metacognitive Awareness: As structured users show behaviors like goal-setting and tracking results, teachers need to provide an explicit instruction of metacognitive strategies when using AI. Students might be asked to keep "AI learning journals" in which they record what AI they used for, what they learned, and how they checked the information. 

4. Implement Multi-Tool Exposure: Due to the heavy reliance on ChatGPT, it is assumed that students would be more prepared if they were given a chance to be systematically exposed to a variety of AI tools (Claude, Gemini, etc.) thus they would acquire a wider digital literacy and be able to comprehend the different capabilities and limitations of ‍‌AI. For Students

The research results provide tips to students of nursing to make use of AI effectively in their learning: 

1. Put in Place a Concrete Plan: Before interacting with AI, students ought to define their learning objectives, regularly use AI in their study sessions, and monitor the impact of AI on their learning. The data indicate that such a structured approach is linked to higher academic self-reported performance. 

2. Employ AI in a Wide-ranging yet Critical Manner: Users with a structured approach interacted with AI in diverse ways (learning concepts, exam preparing, case studying, etc.) and consequently experienced better results. Nevertheless, students should not relinquish critical thinking skills as they must be aware that AI tools may give incorrect information. 

3. Keep Using Traditional Methods along with AI: The conclusion that structured users study for the same or even longer periods of time as before indicates that AI should be the means to traditional learning rather than the end, e.g., reading textbooks, going to lectures, and clinical ‍ ‌practice.

4. ‍Develop Information Verification Habits: More than 70% of students face the problem of AI providing them with wrong information on a daily basis. Therefore, students need to develop disciplined habits of checking AI answers by consulting authoritative nursing sources, textbooks, and evidence-based practice ‍ ‌guidelines. 

For Educational Institutions 

1. Policy Development: Instead of implementing blanket bans on AI, schools should establish detailed policies that see AI as a tool for learning that still needs some level of control and regulation by staff. These policies should identify the educational use of AI as well as the integrity violations that may arise from it. 

2. Helping Teachers: Almalki et al. said in 2025 that nurse teachers see AI differently. Schools should give teachers chances to learn about what AI can do, where it falls short, and how to use it in their teaching.

3. Assessment Redesign: Traditional assessment methods may need redesign in the AI era. The current findings suggest AI helps students understand concepts and prepare for exams—assessment strategies should focus on higher-order thinking skills, clinical application, and ethical reasoning that require genuine understanding beyond AI-generated responses. 

.4. Infrastructure and Access: The provision of equitable access to AI tools and technology services, especially in light of the concerns about the digital divide raised by Nashwan et al. (2025), should be an institutional commitment. 

Study Limitations 

Several limitations should be kept in mind while considering these findings: 

1. Cross-Sectional Design 

This study's cross-sectional design limits the possibility of drawing causal inferences. The associations between structured AI use and self-reported outcomes are positive; however, the authors cannot figure out whether: 

• Structured AI use leads to better performance 

• Students who already perform well are the ones who most likely use AI in a structured manner 

Following students' AI habits and grades as time passes would give us better proof of cause and effect.

2. Self-Reported Data

All variables for academic performance impact, AI effectiveness, and learning outcomes were self-reported rather than objectively measured. Self-report bias can inflate students' ratings of the benefits of AI. Although GPA was self-reported as a categorical variable, this study did not validate GPAs against official academic records nor use objective standardized test scores as measures of performance.

Future research should include objective academic outcomes, such as verified GPAs and standardized exam scores, clinical evaluation results, to complement self-reported perceptions.

3. Convenience Sampling

First, the generalization is limited by the use of convenience sampling. The students who took part in this survey may systematically differ from the non-respondents in their motivation, patterns of AI use, and/or academic performances. This sample over-represents those students comfortable with technology and ready to report on their AI use.

4. Single Institution and Cultural Context

Data have been collected from only one nursing program in the UAE. Findings cannot be generalized to other institutions, countries, and cultural contexts. As Almalki et al. (2025) illustrated, there are significant cross-country differences regarding AI perceptions and usage. The UAE is a country with highly advanced technology infrastructure and educational settings, which may not be representative of other conditions.

5. Classification Methodology

While the classification into structured versus limited AI users was based on theoretically informed criteria, the specific cut-offs (e.g., ≥5 times per week) were somewhat arbitrary. Alternative classification schemes might yield different results. Furthermore, this classification relies on self-reported behaviors, which cannot fully capture the nuances of how students actually integrate AI into their learning.

6. Limited Scope of AI Tools Examined

The review focused on generative AI tools such as ChatGPT, Claude, and Gemini. It did not look at other applications of AI in nursing education such as virtual simulation systems, intelligent tutoring systems, or clinical decision support tools that Shafi et al. (2025) and Nashwan et al. (2025) identified as being key to nursing competency development.

7. Short-Term Outcomes

The present study was therefore limited to measuring immediate perceptions and current academic status but not long-term outcomes related to retention of knowledge, clinical competency development, and performance in professional practice. A systematic review by Shafi et al. (2025) indicated that most research on AI reports short-term gains but lacks substantial evidence on longer-term effects.

8. Potential Confounding Variables

The study did not control for possible confounding variables, for example:

• Prior academic achievement before AI adoption learning style preferences

• Access to other educational resources

•Quality of instruction received

• Clinical placement experiences

• Various socioeconomic factors influence time and resources available for studying.

These variables could affect both the pattern of AI use and academic outcomes.

9. Validity and Reliability Concerns

Although the survey had adapted items from previously validated instruments, formal psychometric testing could not be conducted owing to resource constraints typical in undergraduate research. Internal consistency and factorial structure of the instrument remain unverified.

10. Response Bias

Students might have given socially desirable answers, such as overreporting positive experiences with AI or underreporting concerns about academic integrity, especially if they believe faculty might see the results. The anonymous nature of the survey should have mitigated but may not have eliminated this bias.

Conclusion

This comparative cross-sectional study suggests that more structured use of AI relates to better self-reported academic outcomes for nursing students compared to limited use of AI. Students who reported using AI tools five or more times per week and with a clear, purposeful structure for use, and self-track of outcomes, reported significantly higher rates of improved understanding of concepts (92.3% vs 62.5%), exam confidence (86.5% vs 56.2%), and critical thinking (76.9% vs 50.0%) when compared to students with limited use patterns.

These poductioned outcomes may suggest that quality and intent in AI integration - rather than whether AI exists or is limited - is the critical factor for educational outcomes. Structured users took a comprehensive approach to total AI usage purposes, such as clarifying concepts, exam study preparation, case analysis, and study planning, while systematically recognizing AI usage limitations related to accuracy. 

ChatGPT as the most extensive AI tool and learning resource (used by 90% of structured users) relates to the current market, but indicates a wider possibility for AI literacy in nursing education and practice programs. Students expressed similar dependency and accuracy concerns in both structured and limited use groups, which implies a cognitive awareness of AI limitations related to accuracy across the spectrum of usage. 

The cross-sectional study design limits causal inference; however, the relationship between structured use of AI and positive academic indicators (notably, a higher proportion of structured users received an A-range GPA) suggests that nursing education should focus not on banning AI as an academic aid but on teaching students how to be structured users of AI. The fact that structured users have the same or more study hours than typical study hours suggests that AI is a resource for enhancing learning and not simply a means of cutting corners.

Recommendations 

In light of the above, we recommend the following points:

For Educators:

1. Create an official AI literacy curricula to support students and faculty

2. Create protocols to differentiate between learning uses of AI and academic dishonesty uses

3. Provide faculty professional development that supports pedagogical development for AI usage

4. Develop assessments that challenge higher-order thinking that are difficult to simply ask AI

5. Assess and mitigate access inequities to tech infrastructure that supports and encourages AI

For Nursing Students:

1. Use structured methods to incorporate AI with clear learning objectives to achieve outcomes 

2. Introduce AI across a series of multiple learning events instead of sporadically throughout the semester

3. Begin to develop a habit of verifying information from AI with evidence

4. Begin to track and reflect on the degree and manner in which AI assisted or hindered learning outcomes 

5. Balance the incorporation of AI into study habits and practice with a range of traditional practices and methods of self-directed learning

For Future Research:

1. Longitudinal studies should be conducted to evaluate the impact of AI use and subsequent grades over time

2. Objective evaluation of performance (verified GPAs, standardized exam scores, clinical evaluative frameworks) should be included as outcomes measurement

3. Potential long-term outcomes that could be measured include knowledge retention and competency in clinical settings

4. Mechanisms for structured learning with the incorporated use of AI should be explored for learning

5. Variations of AI tools (ChatGPT, Claude, Gemini, etc.) should be tested to assess whether any provide better educational benefit in terms of student's academic or clinical performance

6. Studies should be developed to assess the relationship between AI and the development of clinical reasoning

7. Studies should be conducted to evaluate AI integration in education as it contributes to preparation for practice and workplace performance and experiences.

Plan of Dissemination and Costing

Dissemination Strategy

The results of the given study will be distributed via several platforms to address the concerned parties in the nursing teaching:

Academic Presentations

• Posting at institutional research symposium or capstone project presentation.

  
  

• Presentation to local nursing education conferences.

  
  

• Presentation of students posters.

  
  

Written Dissemination

• Finally, research report was presented to faculty supervisors and in the institutional library.

  
  

• Summary report given to the administration of the nursing program.

  
  

• Possible submission to undergraduate research journals or nursing education journals.

  
  

Stakeholder Engagement

• The review of implications on curriculum development to nursing faculty.

  
  

• Student forums on the strategies of using AI effectively.

  
  

• Education policy brief on technology and institutional academic integrity committees.

  
  

Timeline

Study Budget & Resources

This research project incurred no direct costs, as all resources were provided through existing institutional infrastructure and digital platforms.

• Survey Platform: Free (Microsoft Forms provided by FCHS)

  
  

• Data Analysis Tools: Free (Excel/SPSS available through the college)

  
  

• Printing or Materials: Not required (all submissions and communications were digital)

  
  

• Meetings and Coordination: Conducted online at no cost

  
  

• Total Project Cost: 0 AED

  
  

Data Management and Intellectual Property

• Any data will be stored in a safe establishment of 5 years in line with the institutional policy and then safely destroyed.

  
  

• Upon a reasonable request and approval, de-identified data can be released to secondary analysis.

  
  

• The results will be disseminated freely without harming the confidentiality of participants.

  
  

• All stakeholders will be well attributed.

References:

Ahmed, F. R., Rushdan, E. E., Al-Yateem, N., Almaazmi, A. N., Subu, M. A., Hijazi, H., Abdelbasset, W. K., Mottershead, R., Ahmed, A. A., & Aburuz, M. E. (2024). AI in higher education: Unveiling nursing students' perspectives on ChatGPT's challenges and opportunities. Teaching and Learning in Nursing. https://doi.org/10.1016/j.teln.2024.12.002

Almalki, M., Alanazi, A., Alshammari, F., Alshammari, M., Almutairi, A., Almutairi, R., Alshammari, M., Alosaimi, H., Alanazi, S., & Alshammari, A. (2025). Integration of artificial intelligence in nursing education: A cross-national exploration. BMC Nursing, 24(1026). https://doi.org/10.1186/s12912-025-01026-0

Al Omari, O., Al Sabei, S., Al-Rawajfah, O., Sharour, L. A., Alrimawi, I., Al Qadire, M., Alabdullah, A., Abu Sharour, L., Halm, M., Bloomer, M. J., Sim, J., & Lopez, V. (2024). Demographic factors, knowledge, attitude and perception and their association with nursing students' intention to use artificial intelligence (AI): A multicentre survey across 10 Arab countries. BMC Medical Education, 24(1456). https://doi.org/10.1186/s12909-024-06447-3

González, L. (2023). Nursing education in the era of ChatGPT: Implications and opportunities. Journal of Nursing Education, 62(9), 493-495. https://doi.org/10.3928/01484834-20230809-01

González-García, A., Bermejo-Martínez, D., López-Alonso, A. I., Treviñón-Redondo, B., Martín-Vázquez, C., & Pérez-González, S. (2024). Impact of ChatGPT usage on nursing students education: A cross-sectional study. Heliyon, 11(1), e41559. https://doi.org/10.1016/j.heliyon.2024.e41559

Khlalf, M. K., Abusamak, M. O., Al-Rousan, M. R., Khasawneh, M. A., & Khlaif, Z. N. (2025). Using generative AI in nursing education: Students' perceptions. BMC Medical Education, 25(926). https://doi.org/10.1186/s12909-025-06926-4

Lei, P., Duo, E. N., Gao, Y., & Zhu, X. (2025). Research on the experience of utilizing artificial intelligence in nursing learning from the perspective of nursing students—A qualitative study. BMC Nursing, 24(1062). https://doi.org/10.1186/s12912-025-01062-w

Nashwan, A. J., Cabrega, J. C. A., Othman, M. I., Khedr, M. A., Osman, Y. M., El-Ashry, A. M., Naif, R., & Mousa, A. A. (2025). The evolving role of nursing informatics in the era of artificial intelligence. International Nursing Review, 72, e13084. https://doi.org/10.1111/inr.13084

Shafi, Z., Samar, V., Rasool, S. U. G., Chapsi, A., & Chandani, S. (2025). Artificial intelligence as an innovative teaching and learning strategy to enhance critical thinking and professional competency in nursing students: A systematic review. Frontiers in Medical & Health Research, 3(3), 782-788.

Su, M., Lin, L., Lin, L., & Chen, Y. (2024). Assessing question characteristic influences on ChatGPT's performance and response-explanation consistency: Insights from Taiwan's Nursing Licensing Exam. International Journal of Nursing Studies, 153, 104717. https://doi.org/10.1016/j.ijnurstu.2024.104717

Taskıran, N. (2023). Effect of artificial intelligence course in nursing on students' medical artificial intelligence readiness: A comparative quasi-experimental study. Nurse Educator, 48(5), E147-E152. https://doi.org/10.1097/NNE.0000000000001452