Introduction
Transformations in the maternal and neonatal health care system are presently occurring swiftly, imposing new requirements on midwifery education.1,2 Midwifery education has evolved into a domain for cultivating professionals equipped with sensitivity, empathy, and a contextual comprehension of the culture and needs of each individual client, rather than merely a venue for acquiring theoretical knowledge or technical abilities.3 As the healthcare system transitions to a value-based care paradigm that prioritizes service value and patient focus, education must evolve accordingly.4 The curriculum is now required to be competency-based, while assessment is transitioning to an outcome-based approach that prioritizes tangible results. Within this framework, success is evaluated not alone by the mastery of theory, but also by the acuity of practice, the sophistication of professional attitudes, and the emotional competencies of students that can be objectively quantified.5,6 Outcome-Based Assessment (OBA) emphasizes the achievement of tangible and measurable outcomes, focusing on competencies that can be directly applied in the workplace, such as technical skills, clinical decision-making, and patient interactions.7
Advancements in Artificial Intelligence (AI) and deep learning are revolutionizing teaching and learning, transforming not just the tools used, but also the educational and developmental processes of midwifery students.8 From the capacity to customize education to the development of authentic clinical simulations,9 AI enhances decision-making by analyzing vast data, enabling adaptive and relevant learning. Key skills like clinical decision-making, once gained through experience, can now be developed using digital methods, especially in resource-limited areas.10
Numerous research indicate that AI can deliver real-time feedback,6 tailored to the needs of each student, so they can learn from mistakes firsthand and develop essential clinical skills.11–13 Convolutional Neural Networks (CNN) are among the technologies that have been employed to integrate various categories of data in order to generate more objective and precise evaluations.14 With the rise of hybrid learning models post-COVID-19, AI plays a key role in creating a more adaptable and personalized learning experience.15 AI systems use algorithms like CNN and Long Short-Term Memory (LSTM) to tailor learning materials to each student’s style and abilities, creating a more adaptive and effective learning experience.16 LSTM, which is a type of neural network in deep learning, is highly effective for analyzing sequential data, such as vital signs or partograph data used in midwifery, to predict the progression of labor and support more accurate clinical decision-making.17
Not only does AI increase learning effectiveness, but it also plays an important role in increasing student engagement.18 Tailoring learning to individual preferences boosts knowledge retention, deepens understanding of clinical material, and sharpens decision-making skills.19,20 For example, CNN and LSTM algorithms enable automated image interpretation and sequential data analysis, both of which are valuable in simulating clinical decision-making processes. In daily midwifery education practice, CNN can be applied in ultrasound image classification to help students identify fetal positions or anomalies, while LSTM support labor progression prediction based on sequential vital signs or partograph data.21 Moreover, in Indonesia, AI adoption in midwifery education is gaining traction, but challenges such as digital literacy gaps, limited resources, and the lack of clear regulations at the local level hinder its full integration into the curriculum. Despite these challenges, AI holds promise in enhancing both the teaching and assessment of clinical competencies, offering a powerful tool for improving the quality of maternal health services in regions where resources are limited.22
Learning chatbots provide real-time feedback and facilitate interactive, question-based engagement, such as simulating patient interviews or guiding students through clinical scenarios step-by-step, which enhances diagnostic reasoning.23 Similarly, Virtual Reality (VR) based midwifery simulation platforms allow students to virtually manage obstetric emergencies like postpartum hemorrhage or shoulder dystocia in safe, immersive environment.24 These tools personalize instruction based on student learning styles and progress, leading to better knowledge retention, deeper understanding of clinical content, and stronger decision-making skills than traditional methods such as lectures or manual Objective Structured Clinical Examinations (OSCEs).25 Furthermore, the application of AI in outcome-based assessment allows for more precise competency evaluations aligned with real-world practice, ultimately contributing to improvements in the quality of maternal health services.26
AI use in midwifery education assessments is not yet optimal, with many institutions using it separately rather than as part of an integrated system. Most studies focus on technical aspects or short-term results, while holistic approaches integrating pedagogy and technology remain scarce.27 There is an urgent need to explore how AI can be used more strategically not just as an aid, but as part of an assessment system that shapes competencies in a sustainable and comprehensive manner.28
AI in OSCE enhances objective and efficient clinical assessments by analyzing student performance in real-time, including movements and communication. It offers quick, personalized feedback and reduces anxiety. ChatGPT aids self-paced exercises and scenario practice. However, ethical concerns like algorithmic bias and misuse must still be addressed.25,28,29
This study aims to provide an original contribution through an in-depth literature review on the role of AI and deep learning in outcome-based assessments in midwifery education. Unlike previous approaches that separate technology and pedagogy, we propose an integrative framework combining technology, assessment, and pedagogy. This study is expected to guide lecturers, curriculum developers, and researchers in designing adaptive, measurable assessment strategies that align with future needs. Additionally, it supports the development of a predictive evaluation system based on student performance, in line with outcome-based education. The results aim to strengthen midwifery graduates’ readiness for clinical challenges and contribute to reducing maternal mortality through responsive, integrity-based, and technology-driven practices.
Materials and Methods
Search Strategy
This study uses literature review methodology. In this literature review, we conducted a search in four databases, namely Taylor and Francis, Science Direct, Semantic Scholar and Springer Nature Link in March 2025, in compliance with the PRISMA 2020 guidelines. Table 1 presents the search terms (keywords) used for literature retrieval across the respective databases. Each database integrates a combination of keywords tailored to the research topic, which focuses on the application of deep learning algorithms in midwifery education and outcome-based assessment.
 |
Table 1 Search Keywords Based on Database
|
A total of 771 articles were identified through a systematic search across four major databases: Taylor and Francis (356 articles), ScienceDirect (27 articles), Semantic Scholar (351 articles), and Springer Nature Link (37 articles). A rigorous screening process was conducted using Rayyan software, which included the removal of duplicate records and an initial selection based on titles and abstracts. Subsequently, two independent reviewers performed a full-text assessment to evaluate each study’s eligibility according to predefined inclusion and exclusion criteria, resulting in 15 articles selected for further analysis.
The data analysis was conducted using a narrative synthesis approach. Two independent reviewers extracted key information from each study, including research design, types of AI or deep learning algorithms employed, educational focus within the context of midwifery education, and the nature and outcomes of the assessments. The findings were then categorized thematically through iterative discussion, highlighting how artificial intelligence contributes to enhancing objectivity, personalized learning, and clinical readiness in outcome-based assessments.
The analysis revealed that algorithms such as CNN, LSTM, Random Forest, and learning chatbots have strong applicability in midwifery education. These technologies facilitate clinical simulation, deliver real-time feedback, and enable more objective and competency-based evaluations. The entire process adhered to the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), and the article selection process is illustrated in the PRISMA flow diagram (Figure 1).30
 |
Figure 1 Flowchart of Study Selection Process for Literature Review on the Integration of Deep Learning in Midwifery Education.
|
Eligibility Criteria
The research to be included in this review will include qualitative and quantitative studies, such as experiments, quasi-experiments, surveys, or cross-sectional studies that address the application of AI and deep learning in midwifery education or that relate to health education that can be applied in midwifery, Table 2 describes the criteria for selecting articles used in the literature selection process. Especially in the context of results-based assessment. In addition, research using a qualitative approach that explores the experiences of students or teachers related to the use of AI and deep learning in midwifery education will also be considered. Studies involving midwifery students enrolled in midwifery education programs or related programs at accredited institutions, either at the bachelor’s, master’s, or diploma levels, will be included, especially if the research addresses how AI and deep learning can improve clinical skills, clinical decision-making, and learning outcomes of midwifery students. The focus of this research is on AI or deep learning technologies, such as simulation-based learning models, gamification, or Natural Language Processing (NLP), as well as how these technologies play a role in personalizing the learning experience and improving the clinical competence of midwifery students. Studies published in the last five years (2020–2025) will be prioritized to ensure the relevance of the latest technologies in midwifery education, by including only research that is freely accessible or that has full access at no cost.
 |
Table 2 Inclusion and Exclusion Criteria
|
Instead, studies that do not use AI or deep learning and cannot be applied in midwifery education will be excluded from this review. Research that only addresses technology-based learning without involving deep learning or AI algorithms, such as online learning platforms or video tutorials that do not use AI elements, will also be excluded. In addition, research that does not compare AI-based learning methods with traditional methods in midwifery education, or that does not examine the impact of technology on the practical skills or decision-making of midwifery students, will be ignored. Studies involving experienced health practitioners or professionals will also be excluded without involving relevant midwifery students or health education students. Research published before 2020 will be reviewed to ascertain the relevance of the technology used in the study to the latest developments in midwifery education. Only studies that are freely accessible or have a full-access version at no cost will be considered, ensuring that all selected literature is accessible to general readers without cost barriers.
Results
Overview of Included Studies
A total of 15 articles included in this review are summarized in Table 3, which presents an overview of the deep learning algorithms employed in each study published between 2020 and 2025. The analysis of publication year characteristics indicates that most studies were published in 2025, representing seven articles, which accounts for 47% of the total. The year 2022 ranks second with four publications (27%), while 2023 and 2024 each contributed two articles (13%). This distribution reflects a significant increase in publication volume in recent years, particularly in 2025, which indicates growing academic attention and shifting research focus within the field. The analyzed references provide a detailed account of annual contributions, underscoring the dynamic and continuously expanding nature of scholarly discourse in this domain. These details are comprehensively presented in Figure 2: Analysis of Publication Year Characteristics in Studies from 2022 to 2025. These studies came from 12 countries, such as China,31–34 South Korea,23,35 Iran,36 Ghana,37 Bulgaria,38 Australia,39 Indonesia,40 India,41 Syria,42 Estonia,35 United States,35,43 and Greece,44 which shows that the adoption of artificial intelligence (AI) in midwifery education is a global issue that crosses cultural boundaries and education systems. Relevant data are presented Figure 3, which illustrates the distribution of studies based on their country of origin and corresponding study references.
 |
Table 3 Summary of Research on the Application of Deep Learning Algorithms in Midwifery Education (2021–2025)
|
 |
Figure 2 Analysis of Publication Year Characteristics in Studies from 2020 to 2025.
|
 |
Figure 3 Distribution of Studies Based on Country of Origin and Study References.
|
The primary focus of these studies is midwifery education as many as eight studies,31,32,36–39,41,44 teaching evaluation four studies,33,35,39,43 and the use of AI in clinical decision-making or pregnancy prediction two studies.32,43 The data on the distribution of studies based on the primary research objectives is illustrated through the table representation in Table 4.
 |
Table 4 Distribution of Studies Based on the Main Research Objectives
|
In research in the field of Obstetrics Education and the Use of AI for Pregnancy Prediction, the cross-sectional study design of 3 studies31,37,42 and a retrospective 1 study32 is used to analyze phenomena at a specific point in time or look at patterns in historical data. To develop critical thinking skills in midwifery students, a randomized controlled trial crossover 1 study36 is an effective option, allowing researchers to test the changes that occurred before and after the intervention. On the other hand, in the Evaluation of Teaching and the Use of AI in Clinical Decision Making34 The design of the experimental study there are 6 studies23,33,38,44,45 and simulation-based experiments there was 1 study35 is widely applied to test new methods or technologies that can improve teaching outcomes and support data-driven decision-making. Finally, a quantitative study with time series forecasting 1 study,41 is used in mobile app-based Midwifery Education to analyze user interaction data and personalize students’ learning experiences according to their needs. Figure 4 provides a detailed depiction of the diverse research designs utilized throughout this study.
 |
Figure 4 Distribution of Studies According to Research Design.
|
Based on the analysis of 15 studies, deep learning algorithms applied in the context of midwifery education can be classified into three main categories. First, CNN and its derivatives are widely used for visual analysis, such as in facial expression recognition to improve students’ nonverbal communication,31 prediction of the success of the birth from the image of the embryo,32 and evaluation of the effectiveness of teaching based on learning data.33 Second, the Deep Neural Networks (DNN) algorithm, including its variations such as Multilayer Perceptron (MLP) and Fully Connected Neural Networks (FCNN), was used in research related to the prediction of learning interest in hybrid systems,40 integration of symbolic knowledge in AI learning,35 and maternal health risk classification.44 Third, types of Recurrent Neural Networks (RNNs), such as LSTM and Deep AR, are applied to evaluate factors that affect the quality of teaching,34 as well as in predicting the demand for learning content in real-time-based midwifery applications.41 In addition, some studies have chosen general AI-based approaches or non-deep learning machine learning such as Random Forest and Support Vector Machine (SVM), or even do not explicitly mention the type of algorithm used. Table 5 provides a detailed representation of the various deep learning algorithms utilized within the scope of this study. These findings show that deep learning algorithms are increasingly growing in supporting innovation and effectiveness of midwifery learning in the digital era.
 |
Table 5 Deep Learning Algorithm Types and Their Roles in Midwifery Education Research
|
AI Application Across Midwifery Competency Domains
AI has been applied to support the development of all domains of obstetric competencies, namely cognitive, affective, and psychomotor. In the cognitive domain, technologies such as Random Forest and LSTM are used to predict pregnancy risk and maternal complications.32,44 These models help students understand complex clinical patterns through data-driven analysis. In the affective domain, AI approaches are used to develop emotional and interpersonal aspects, for example with a CNN-based facial recognition system that has been shown to improve the emotional awareness and nonverbal communication skills of midwifery students.31 Additionally, the use of technology-based reflective learning allows students to evaluate their learning process in a more in-depth and evidence-based manner.39 In the psychomotor realm, technologies such as VR and AI simulation are used to improve students’ technical and procedural skills, providing a safe and realistic practical experience.38
Types and Functions of AI Technologies Used
The variety of AI technologies used in these studies is quite wide, including CNN algorithms, Random Forest, LSTM, Deep Neural Networks (DNN), Neural-Symbolic AI (NSAI) approaches, and NLP-based chatbots. Its functions are also diverse, ranging from classification and clinical prediction,32,44 evaluation of learning with fuzzy logic and CNN models,34 to text-based interaction through chatbots for Electronic Fetal Monitoring (EFM) training.23 Other studies show the use of AI for personalized learning based on user behavior patterns, such as in DeepAR-based mobile apps.41 The implementation of NSAI is also an important breakthrough in increasing interpretability and transparency in the education decision support system. AI technology in general has served not only to speed up the decision-making process, but also to create a more responsive and adaptive learning experience.35
Pedagogical Frameworks Supporting AI Integration
The integration of AI in midwifery education goes hand in hand with the application of modern pedagogical approaches such as experiential learning, reflective practice, and constructivism. For example, experiential learning is realized through the use of VR and interactive simulations that allow students to experience clinical scenarios firsthand without real risks.38 Meanwhile, reflective practice is supported by multimodal learning analytics (MMLA) technology, which provides data-driven feedback that helps students evaluate their own learning process.39 A constructivist approach is also seen in the use of WebQuest and hybrid learning, where students actively build understanding through exploration with the guidance of individually tailored AI systems. Thus, AI becomes an integral part of shaping a meaningful and contextual learning ecosystem.36 A synthesis of the findings of previous studies is presented in the form of a diagram in Figure 5.
 |
Figure 5 AI Enhances Midwifery education.
|
The diagram illustrates the application of artificial intelligence (AI) in midwifery education, focusing on three main competency domains: affective, cognitive, and psychomotor. Technologies such as NLP-based chatbots, deep learning algorithms, and clinical prediction models play an important role in supporting more personalized learning, data-driven decision-making, and safer and more realistic practical skills training. Deep learning algorithms analyze complex data to uncover deep patterns, enriching the learning experience. This diagram shows how innovative technology can create more adaptive and meaningful midwifery learning.
Discussion
Transformation of Midwifery Education Through AI and Deep Learning
The incorporation of AI and deep learning algorithms in midwifery education is becoming a transformative influence that improves multiple areas of student proficiency, including cognitive, affective, and psychomotor dimensions46–48 The fifteen studies included in this literature review provide a comprehensive overview of the diverse applications of AI technology, such as pedagogical evaluation, curriculum personalization, clinical decision-making support, and enhanced communication.
Investigate innovative applications of CNN to develop models proficient at recognizing facial expressions. This strategy seeks to assist nursing students in enhancing their nonverbal communication skills, a facet that is frequently challenging to impart through traditional methods. CNN, proficient in visual data processing, can identify intricate patterns of emotional expression and deliver real-time feedback during student interactions in clinical simulations. In midwifery education, interpreting and reacting to a client’s body language transcends mere technique; it fundamentally involves cultivating empathy and trust.31 This aligns with Burgoon and Hale’s theory of interpersonal communication, which posits that most emotional messages are transmitted through nonverbal language.49–51 This approach also supports Kolb’s theory of experiential learning, where the most effective learning occurs through hands-on experience. With CNN technology, students not only practice, but can also adapt their communication style based on the feedback received, so that interpersonal skills can be honed in a more in-depth and sustainable manner.52–54
Clinical Decision Support and Contextual Learning
Conversely, a comparable technique, specifically CNN, is employed to analyze photos of embryos to predict live birth outcomes with remarkable precision. This method provides fresh insights into how AI might enhance clinical decision-making, especially in the reproductive domain, which is replete with intricate visual data that is challenging to comprehend manually.32 From an educational standpoint, this necessitates that midwifery students comprehend both theory and practice, while also acquiring the ability to incorporate new technologies to facilitate evidence-based decision-making. The cognitive apprenticeship hypothesis, which highlights learning through expert mentorship, is especially pertinent in this context, as students are instructed to emulate and comprehend the application of technology by specialists in actual clinical scenarios. AI-enhanced case-based simulations serve as an excellent medium for developing these skills, situating learning within a more contextual and practical framework.9,55
Cognitive Learning Innovation Through AI-Based WebQuest
In the cognitive domain, the WebQuest model is an innovative learning technique, comprising a structured online activity aimed at promoting active exploration and resolution of clinical problems by students. In health education, AI-based WebQuest enables students to engage in contextual and collaborative learning, addressing genuine clinical difficulties with advanced analytical methods. The implementation of this technology has demonstrated efficacy in enhancing student engagement, fortifying critical thinking abilities, and cultivating academic self-efficacy.36
Data-Driven Learning and Teaching Evaluation
The evaluation methodology utilizes CNN in conjunction with data mining techniques to examine diverse educational data sources, including student feedback, lesson planning, classroom interactions, and teacher performance. This system is executed as a digital platform capable of real-time data collection and processing, subsequently offering an interactive dashboard for lecturers as a means of evaluation and instructional reflection.33
This feature enables teachers to proactively modify learning strategies based on student needs and recent advancements in clinical practice. This evaluation, grounded in data, enhances objectivity in teaching quality assessment and facilitates ongoing professional development for lecturers, leading to notable improvements in students’ practical skills.
Data-driven evaluations enhance objectivity in the assessment of teaching quality and directly support the professional development of lecturers as well as the enhancement of students’ practical skills. The study conducted in Ghana presents an innovative approach by utilizing machine learning to evaluate the nursing and midwifery informatics curriculum. This study achieved high accuracy in predicting students’ academic performance, reaching up to 95%, by employing algorithms including Random Forest, Gradient Boosting, SVM, and Logistic Regression. This model has enhanced student engagement in the learning process, representing a significant achievement in the context of global challenges, including resource limitations and a substantial maternal health burden. To implement this approach effectively, educational institutions must possess adequate digital infrastructure and strategies to address the data literacy gap among both lecturers and students.37
The establishment of a teaching quality assessment system through the integration of fuzzy logic and LSTM algorithms within the same framework. This combination demonstrated the capability to yield a very accurate rating of 98.4% in identifying the primary factors influencing teaching quality. This method facilitates the development of learning modalities in midwifery education that are more attuned to student needs. The successful application of this technology is contingent upon the preparedness of institutions to offer technical training and establish trustworthy data management infrastructure, particularly in nations with restricted internet access.34,56
The NSAI approach enhances the learning assessment perspective by integrating the data processing capabilities of deep neural networks with the precision of symbolic knowledge-based reasoning. The primary benefit of this method is its capacity to deliver predictions of academic achievement that are both precise and logically comprehensible. NSAI is highly pertinent in obstetric education, particularly for clinical decision-making that necessitates high accountability and openness.35 Nonetheless, the incorporation of this technology continues to encounter obstacles, including the intricacy of system architecture and the requirement for substantial computational resources. Consequently, a strong partnership among educational practitioners, technology specialists, and policymakers is essential to guarantee the inclusive adoption of these technologies and their significant influence on future midwifery learning environments.57
In a global context influenced by the pandemic, accelerating digitalization, and mobility of health workers, the integration of technology in learning and evaluation such as AI-based WebQuest and CNN-based evaluation has become particularly relevant. Both maintain the continuity of quality learning in crisis situations, and also become an important foundation in shaping future health workers who are resilient, critical, and able to face multidimensional challenges ranging from climate change and global health crises, to service complexities due to urbanization and globalization.48
AI and VR-Based Simulations for Psychomotor Skill Development
Aligned with the emphasis on assessing and enhancing learning quality, the research conducted by Aabaah, Yu, and Hooshyar has further advanced the application of deep learning technology in midwifery education. The integration of VR technology with AI has been recognized as a significant advancement in facilitating the development of students’ psychomotor abilities in the health sector, especially in midwifery education. Utilizing an innovative, immersive, and secure simulation-based learning methodology, it enables students to engage in many clinical procedures within a three-dimensional virtual environment that closely mirrors actual practice.37 Moreover, AI integration enables the system to deliver automatic feedback, perpetually monitor user performance, and adaptively modify the exercise’s complexity according to individual progress. The integration of VR and AI establishes a learning environment that facilitates repetitive practice without jeopardizing patient safety, simultaneously enhancing students’ clinical proficiency and professional preparedness prior to their entry into practical settings.38
This strategy addresses the growing complexity of global concerns, particularly the restricted access to clinical practices intensified by the COVID-19 pandemic. As hospitals and health facilities restrict student involvement, the necessity for alternate educational resources that uphold the quality of training becomes increasingly imperative. Despite its potential, the deployment of this technology is fraught with numerous challenges58,59 Substantial investment expenditures for hardware acquisition and software development continue to be a significant obstacle, particularly for institutions with constrained resources. Moreover, differences in digital literacy between educators and students, opposition to the adoption of long-standing instructional methodologies, and concerns over ethics and data protection constitute problems necessitating a holistic management plan.60–63
This approach, which is based on VR and AI, not only resolves technical issues in the learning of clinical skills, but also makes a strategic contribution to the endeavor to prepare healthcare workers across countries and adapt to international competency standards in an increasingly digitized global education landscape.64 The ability to create a flexible and adaptive learning ecosystem renders it pertinent in the context of transnational challenges, including rapid urbanization, climate change, and the global mobility of health workers. This technology integration should be regarded not merely as a temporary remedy for existing practice limitations, but as a fundamental component of the long-term evolution of midwifery education towards a more inclusive, responsive, and evidence-based framework.65
Adaptive Learning and Chatbot Models for Learning Autonomy
The digital change in midwifery education has markedly increased with the integration of artificial intelligence (AI) and data-driven learning tools, particularly with the transition to remote and hybrid learning models prompted by the COVID-19 pandemic. The creation of a chatbot-based educational model utilizing NLP to enhance students’ motivation and learning autonomy on EFM. The chatbot operates as a virtual instructor, delivering real-time explanations, contextually answering queries, and offering individualized emotional support, so enhancing the human learning experience in an online environment. The significance of this strategy in midwifery education is considerable, as subjects like cardiotocography interpretation (CTG) necessitate a profound clinical comprehension that is challenging to attain without continuous mentorship.23,66
This approach employs multimodal learning analytics to enhance collaborative learning by analyzing students’ digital interactions encompassing text, audio, and visual elements. This method offers predictive and formative feedback that enhances the reflection and modification of student learning strategies. In midwifery education, where teamwork and data-informed decision-making are crucial, these tools facilitate the development of vital professional qualities.39 Research indicates that academic satisfaction is a crucial factor in sustaining students’ engagement in hybrid learning. These findings validate that good and significant learning experiences can enhance the sustainability of digital learning models among midwifery students worldwide.40
Ethical Challenges, Digital Literacy, and Infrastructure Gaps
The integration of AI in education is inherently linked to several issues. Capable of recognizing deficiencies in AI literacy among educators and students as the primary obstacles affecting the efficacy of technology implementation.42 Research by Guitart et al indicates that predictive analytics in mobile learning applications possesses significant potential for material customization; nevertheless, its efficacy is substantially affected by the user’s comprehension of algorithmic input. Additional obstacles encompass ethical considerations and data protection concerns.41 Emphasize the significance of stringent laws to guarantee privacy and responsibility in the application of AI within educational and maternity healthcare environments.45 The ethical significance is enhanced by demonstrating AI’s capacity to identify maternal dangers promptly inside safe and evidence-based clinical decision-making scenarios. This strategy provides an alternative training that is both pragmatic and very contextual in the cultivation of clinical skills for midwifery students.44
Paradigm Shift in Learning Toward Adaptive and Competency-Based Models
These studies highlight a change from conventional teaching methods to more adaptable, reflexive, and competency-based learning. AI serves as a facilitator in assessment for learning, providing feedback that is both formative and capable of guiding students’ sustained self-development. This necessitates a curriculum that clearly incorporates digital competency, technological ethics, and data literacy within the midwifery education framework. This type of integration necessitates institutional preparedness to furnish sufficient infrastructure and professional development for educators.67
Global Implications and Future Research Directions
In a global context, particularly in low and middle income countries confronting significant maternal health challenges and a deficit of educators, AI technology and digital learning analytics can be essential in addressing disparities in educational quality48,68,69 However, equal access to digital technology and infrastructure is an absolute prerequisite to ensure further inclusivity,70 Long-term challenges such as climate change, cross-border migration of health workers, and the globalization of health care systems also reinforce the need for a flexible, data-driven, and systemic disruption-resistant approach to education.59 Consequently, subsequent research ought to concentrate on a longitudinal assessment of the influence of AI integration on the development of professional capabilities among midwifery students, emphasizing algorithm transparency, system accountability, and global technology affordability.
Conclusion
Impact of AI on Midwifery Education
The integration of AI and deep learning algorithms significantly enhances outcome-based assessment in midwifery education. Technologies such as CNN, LSTM, Random Forest, and SVM not only improve assessment objectivity but also personalize learning and better prepare students for clinical practice. These tools foster the development of clinical decision-making, communication skills, and evidence-based reasoning, which are essential for effective midwifery practice.
Challenges in AI Implementation
Despite its potential, challenges such as technological literacy gaps among educators and students, digital infrastructure limitations, and ethical concerns related to data privacy and algorithm transparency remain significant barriers to AI adoption in midwifery education. These issues must be addressed to ensure equitable and effective implementation at all levels.
Recommendations for Educators
To optimize AI’s impact, educators must receive comprehensive training to enhance digital literacy and ethical understanding of AI in midwifery contexts. This training should focus on personalized learning, improving clinical communication, and evidence-based decision-making, ensuring that educators are equipped to leverage AI effectively while addressing ethical concerns.
Recommendations for Educational Institutions
Institutions must invest in robust digital infrastructure to support AI-based tools like CNN and SVM. Clear policies on data privacy and algorithm transparency are also essential to mitigate ethical issues and ensure that AI is used responsibly.
Competency-Based Curriculum Development
Midwifery curricula should be competency-based and tailored to local contexts, addressing the specific needs of students. Collaboration between academic institutions, technology developers, and policymakers is critical to ensure AI implementation is sustainable and inclusive.
Limitations of the Study
This review highlights that many of the studies included focus primarily on the technical aspects of AI, with limited exploration of long-term effects or real-world applicability. There is a need for more research on how AI technologies can be scaled and adapted in diverse educational settings, especially in low-resource environments.
Strengths and Weaknesses
The strengths of AI integration in midwifery education include its potential to provide personalized learning and enhance clinical competencies. However, there is a notable weakness in the ethical considerations surrounding AI, particularly concerning data privacy and transparency, which must be addressed in future developments.
Future Research
Further longitudinal studies are required to assess the long-term effects of AI on midwifery competencies. Additionally, research should focus on developing AI models that are both ethically sound and contextually adaptable, reducing algorithmic bias and enhancing transparency in midwifery education.
Final Conclusion
AI offers significant potential to transform midwifery education, but its integration must be done cautiously and comprehensively. It is essential to maintain social justice and ethical standards in maternal and child health services while improving educational outcomes. The inclusive adoption of AI will not only enhance midwifery education but also contribute to improving maternal health outcomes, particularly in low- and middle-income countries with high maternal and infant mortality rates.
Data Sharing Statement
Data are reported in the current study.
Acknowledgments
The author expresses gratitude to all colleagues whose contributions have facilitated this literature review. We extend our sincere gratitude to the Master of Midwifery Program, Faculty of Medicine, for granting access to invaluable academic resources. Gratitude is extended to the supervisors for their indispensable direction during this review, along with the anonymous reviewers for their helpful criticism that enhanced the quality of this article.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Funding
There is no funding to report.
Disclosure
The authors wish to confirm that there are no conflicts of interest associated with this research.
References
1. Marcotte KM, Gruppen LD. Competency-based education as curriculum and assessment for integrative learning. Educ Sci. 2022;12(4):267. doi:10.3390/educsci12040267
2. Smith RM, Calvert B, Kata E, et al. Midwifery education in Pacific Island countries: a discussion paper. Women Birth. 2023;36(6):e605–12. doi:10.1016/j.wombi.2023.05.008
3. McKellar L, Graham K, Sheehan A, Fleet JA, Sidebotham M, Sweet L. Examining the transformation of midwifery education in Australia to inform future directions: an integrative review. Women Birth. 2023;36(2):155–166. doi:10.1016/j.wombi.2022.11.010.
4. Seyednoori T, Khorsandi Taskoh A, Jamehbozorg Z, Yadegarzadeh G, Abdollahi H. Designing and validating an educational model based on accountable midwifery. Res Med Educ. 2022;14(3):71–85. doi:10.52547/rme.14.3.71
5. Firoozehchian F, Zareiyan A, Geranmayeh M, Behboodi Moghadam Z. Domains of competence in midwifery students: a basis for developing a competence assessment tool for iranian undergraduate midwifery students. BMC Med Educ. 2022;22(1):1–15. doi:10.1186/s12909-022-03759-z
6. Perkins M, Furze L, Roe J, Macvaugh J. The artificial intelligence assessment scale (AIAS): a framework for ethical integration of generative AI in educational assessment. J Univ Teach Learn Pract. 2024;21(6).
7. Agir N, Effendi M, Matore EM, Faamanatu-Eteuati N, Marquez N. Outcome-based assessment in the evaluation of education programs through a systematic literature review. Int J Acad Res Progress Educ Dev. 2023;12(2):2662–2677.
8. Karabacak M, Ozkara BB, Margetis K, Wintermark M, Bisdas S. The advent of generative language models in medical education. JMIR Med Educ. 2023;9.
9. Pajohideh ZS, Mohammadi S, Keshmiri F, Jahangirimehr A, Honarmandpour A. The effects of normal vaginal birth simulation training on the clinical skills of midwifery students: a quasi-experiment study. BMC Med Educ. 2023;23(1):1–7. doi:10.1186/s12909-023-04319-9
10. Kim HY, HSK GJC. Applications of artificial intelligence in radiology. Radiologie. 2023;42:2–9.
11. Bigirwa JP, Ndawula S, Naluwemba EF. Technology leadership practices of end users and the adoption of E-learning in midwifery institutions in Uganda. J Med Educ Curric Dev. 2022;9:238212052210963. doi:10.1177/23821205221096376
12. Gao Y. Deep learning-based strategies for evaluating and enhancing university teaching quality. Comput Educ Artif Intell. 2025;8:100362. doi:10.1016/j.caeai.2025.100362
13. Shah C, Davtyan K, Nasrallah I, Bryan RN, Mohan S. Artificial intelligence-powered clinical decision support and simulation platform for radiology trainee education. J Digit Imaging. 2023;36(1):11–16. doi:10.1007/s10278-022-00713-9
14. Feng G, Fan M, Chen Y. Analysis and prediction of students’ academic performance based on educational data mining. IEEE Access. 2022;10:19558–19571. doi:10.1109/ACCESS.2022.3151652
15. Aguilar-Esteva V. Adán Acosta-Banda RCA and MPO. Sustainable social development through the use of artificial intelligence and data science in education during the COVID emergency_ A systematic review using PRISMA.pdf. 2023; Available from:
16. Osaka J, Maeda A, Oka H, Mori Y, Ishioka T, Suyari H. Reliable and efficient automated short-answer scoring for a large dataset using active learning and deep learning. Interact Learn Environ. 2025;1–12. doi:10.1080/10494820.2025.2452005
17. Aljaloud AS, Uliyan DM, Alkhalil A, et al. A deep learning model to predict student learning outcomes in LMS using CNN and LSTM. IEEE Access. 2022;10:85255–85265. doi:10.1109/ACCESS.2022.3196784
18. Ma Y, Wei Y, Shi Y, Li X, Tian Y, Zhao Z. Online learning engagement recognition using bidirectional long-term recurrent convolutional networks. Sustain. 2023;15(1):1–14.
19. Kivlighan KT, Ortiz F, Migliaccio L, et al. Thirty years of midwifery education in the rural and diverse state of New Mexico. J Midwifery Women’s Heal. 2024;157–162.
20. Ulfa Y, Igarashi Y, Takahata K, Horiuchi S. Effects of team-based learning about postpartum haemorrhage on learning outcomes and experience of midwifery students in Indonesia: a pilot study. Nurs Open. 2021;8(1):241–250. doi:10.1002/nop2.623
21. Kim HY, HSK GJC, Kwon HS. Applications of artificial intelligence in radiology. Ultrasonography. 2023;42(January):2–9.
22. Wadipalapa RP, Katharina R, Nainggolan PP, et al. An ambitious artificial intelligence policy in a decentralised governance system: evidence from Indonesia. J Curr Southeast Asian Aff. 2024;43(1):65–93. doi:10.1177/18681034231226393
23. Han JW, Park J, Lee H. Analysis of the effect of an artificial intelligence chatbot educational program on non-face-to-face classes: a quasi-experimental study. BMC Med Educ. 2022;22(1):1–10. doi:10.1186/s12909-022-03898-3
24. Zhao L, Dai X, Chen S. Effect of the case-based learning method combined with virtual reality simulation technology on midwifery laboratory courses: a quasi-experimental study. Int J Nurs Sci. 2024;11(1):76–82. doi:10.1016/j.ijnss.2023.12.009
25. Misra SM, Suresh S. Artificial intelligence and objective structured clinical examinations: using ChatGPT to revolutionize clinical skills assessment in medical education. J Med Educ Curric Dev. 2024;11(Llm):1–6. doi:10.1177/23821205241263475
26. Nugraheny E, Yanti Y, Joewono HT. Why we have to reform midwifery education? J Pendidik Kedokt Indones Indones J Med Educ. 2022;11(3):332.
27. Kranz A, Abele H. The impact of artificial intelligence (AI) on midwifery education: a scoping review. Healthc. 2024;12(11):1–13. doi:10.3390/healthcare12111082
28. Ogunsakin B, Babatimehin T. Unlocking the potential of artificial intelegence: a new paradigm for assessment in 21 st century eduction. Int. J. Res. STEM Educ. 2024;6(2):37–49.
29. Diekuu JB, Mekala MS, Abonie US, Isaacs J, Elyan E. Predicting student next-term performance in degree programs using AI-based approach: a case study from Ghana. Cogent Educ. 2025;12(1). doi:10.1080/2331186X.2025.2481000
30. Page MJ, Mckenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. 2021;10(1):1–11. doi:10.1186/s13643-021-01626-4
31. Yu IC, Guo JM, Lin WC, Fang JT. Development of nonverbal communication behavior model for nursing students based on deep learning facial expression recognition technology. Cogent Educ. 2025;12(1). doi:10.1080/2331186X.2024.2448059
32. Huang B, Zheng S, Ma B, Yang Y, Zhang S, Jin L. Using deep learning to predict the outcome of live birth from more than 10,000 embryo data. BMC Pregnancy Childbirth. 2022;22(1):1–7. doi:10.1186/s12884-021-04373-5
33. ZHOU S, LI D. Construction of teacher learning evaluation model based on deep learning data mining. Scalable Comput. 2025;26(1):415–422.
34. Yu J. Evaluation of influencing factors of China university teaching quality based on fuzzy logic and deep learning technology. PLoS One. 2024;19(9):1–20. doi:10.1371/journal.pone.0303613
35. Hooshyar D, Azevedo R, Yang Y. Augmenting deep neural networks with symbolic educational knowledge: towards trustworthy and interpretable AI for education. Mach Learn Knowl Extr. 2024;6(1):593–618. doi:10.3390/make6010028
36. Mohammadi A, Modarres M, Khakbazan Z, Sadat Hoseini AS, Asghari Jafarabadi M, Geranmayeh M. Effect of WebQuest‑based education on critical thinking and academic self‑efficacy of midwifery students: study protocol of a randomized, controlled crossover trial. J Educ Health Promot. 2023;12(1). doi:10.4103/jehp.jehp_1662_22
37. Aabaah I, Wiredu JK, Emmanuel B. Revolutionizing nursing and midwifery informatics curriculum evaluation in Ghana: a data-driven machine learning approach. J. Inf. Syst. Inform. 2025;7(1):442–460. doi:10.51519/journalisi.v7i1.1018
38. Georgieva-Tsaneva G, Serbezova I, Beloeva S. Application of virtual reality, artificial intelligence, and other innovative technologies in healthcare education (nursing and midwifery specialties): challenges and strategies. Educ Sci. 2025;15(1):1–23. doi:10.3390/educsci15010011
39. Yan L, Echeverria V, Jin Y, et al. Evidence-based multimodal learning analytics for feedback and reflection in collaborative learning. Br J Educ Technol. 2024;55(5):1900–1925. doi:10.1111/bjet.13498
40. Yaqin AM, ‘Ainul, Muqoffi AK, Rizalmi SR, Pratikno FA, Efranto RY. Hybrid learning in post-pandemic higher education systems: an analysis using SEM and DNN. Cogent Educ. 2025;12(1). doi:10.1080/2331186X.2025.2458930
41. Guitart A, Del Río AF, Periáñez Á, Bellhouse L. Midwifery learning and forecasting: predicting content demand with user-generated logs. Artif Intell Med. 2023;138(November 2021):102511. doi:10.1016/j.artmed.2023.102511
42. Swed S, Alibrahim H, Elkalagi NKH, et al. Knowledge, attitude, and practice of artificial intelligence among doctors and medical students in Syria: a cross-sectional online survey. Front Artif Intell. 2022;5(10):1–15. doi:10.3389/frai.2022.1011524
43. Avishek Choudhury OA. Impact of accountability, training, and human factors on the use of artificial intelligence in healthcare_ Exploring the perceptions of healthcare practitioners in the US. Hum Factors Heal. 2022; 2(100021):1–12. doi:10.1016/j.hfh.2022.100021
44. Tzimourta KD, Tsipouras MG, Angelidis P, Tsalikakis DG, Orovou E. Maternal health risk detection. Adva Midwifery Artifi Intellig. 2025;13(7):1–21. doi:10.3390/healthcare13070833
45. Choudhury A, Asan O. Impact of accountability, training, and human factors on the use of artificial intelligence in healthcare: exploring the perceptions of healthcare practitioners in the US. Hum Factors Healthc. 2022;2(January):100021. doi:10.1016/j.hfh.2022.100021
46. Alqahtani T, Badreldin HA, Alrashed M, et al. The emergent role of artificial intelligence, natural learning processing, and large language models in higher education and research. Res Soc Adm Pharm. 2023;19(8):1236–1242. doi:10.1016/j.sapharm.2023.05.016
47. Ilić M, Mikić V, Kopanja L, Vesin B. Intelligent techniques in e learning: a literature review [Internet]. Artif Intell Rev. 2023. doi:10.1007/s10462-023-10508-1
48. Malerbi FK, Nakayama LF, Dychiao RG, et al. Digital education for the deployment of artificial intelligence in health care. J Med Internet Res. 2023;25:1–4. doi:10.2196/43333
49. Villegas-Ch WE, Garcia-Ortiz J, Sanchez-Viteri S. Identification of emotions from facial gestures in a teaching environment with the use of machine learning techniques. IEEE Access. 2023;11(March):38010–38022. doi:10.1109/ACCESS.2023.3267007
50. Kwon J, Kotani H. Head motion synchrony in unidirectional and bidirectional verbal communication. PLoS One. 2023;18(5 May):1–15. doi:10.1371/journal.pone.0286098
51. Derakhshan A, Zhang LJ, Zhaleh K. The effects of instructor clarity and non-verbal immediacy on Chinese and Iranian EFL students’ affective learning: the mediating role of instructor understanding. Stud Second Lang Learn Teach. 2023;13(1):71–100. doi:10.14746/ssllt.31733
52. Parkinson B. Cross-cultural calibration of words and emotions: referential, constructionist, and pragmatic perspectives. Emot Rev. 2023;15(4):348–362. doi:10.1177/17540739231182680
53. McLachlan K, Tippett N. Kickstarting creative collaboration: placing authentic feedback at the heart of online digital media education. Assess Eval High Educ. 2024;49(2):246–261. doi:10.1080/02602938.2023.2209295
54. Clarke SO, Ilgen JS, Regehr G. Fostering adaptive expertise through simulation. Acad Med. 2023;98(9):994–1001. doi:10.1097/ACM.0000000000005257
55. Janighorban M, Yousefi H, Yamani N. Structural empowerment of midwifery students following simulation-based training of management of obstetric emergencies. BMC Med Educ. 2023;23(1):1–8. doi:10.1186/s12909-023-04365-3
56. Gong T, Wang J. A data-driven smart evaluation framework for teaching effect based on fuzzy comprehensive analysis. IEEE Access. 2023;11(February):23355–23365. doi:10.1109/ACCESS.2023.3253379
57. Dimitriadou E, Lanitis A. A critical evaluation, challenges, and future perspectives of using artificial intelligence and emerging technologies in smart classrooms. Smart Learn Environ. 2023;10(1). doi:10.1186/s40561-023-00231-3
58. García Zare EJ, Soto Abanto SE, Rodriguez Paredes NP, et al. Technological devices and digital competences: a look into the digital divides for university continuity during the COVID-19 pandemic. Sustain. 2023;15(11):1–16. doi:10.3390/su15118494
59. Pancheshnikov A, Cuneo CN, Matias WR, et al. Case studies in adaptation: centring equity in global health education during the COVID-19 pandemic and beyond. BMJ Glob Heal. 2023;8(4):1–10.
60. Hertling SF, Back DA, Kaiser M, Loos FM, Schleußner E, Graul I. Students’ and lecturers’ perspectives on the implementation of online learning in medical education due to COVID-19 in Germany: a cross-sectional pilot study. Front Med. 2023;10(2023):1–9.
61. Atenas J, Havemann L, Timmermann C. Reframing data ethics in research methods education: a pathway to critical data literacy. Int J Educ Technol High Educ. 2023;20(1):1–27. doi:10.1186/s41239-023-00380-y
62. Farias-Gaytan S, Aguaded I, Ramirez-Montoya MS. Digital transformation and digital literacy in the context of complexity within higher education institutions: a systematic literature review. Humanit Soc Sci Commun. 2023;10(1):1–11. doi:10.1057/s41599-023-01875-9
63. Stewart B, Miklas E, Szcyrek S, Le T. Barriers and beliefs: a comparative case study of how university educators understand the datafication of higher education systems. Int J Educ Technol High Educ. 2023;20(1):1–20. doi:10.1186/s41239-023-00402-9
64. Frehywot S, Vovides Y. An equitable and sustainable community of practice framework to address the use of artificial intelligence for global health workforce training. Hum Resour Health. 2023;21(1):1–7. doi:10.1186/s12960-023-00833-5
65. Birkheim SL, Calogiuri G, RM. Advancing immersive virtual reality-based simulation practices: developing an evidence-based and theory-driven pedagogical framework for VR-based simulation. Interact Learn Environ. 2024;32(7):3579–3591. doi:10.1080/10494820.2023.2186896
66. Moroni EA, Kim JS, Chen W, Solari MG, Wachtman G, Manders EK. A novel rib removal technique for thoracic contouring and obtaining rib bone grafts. Plast Reconstr Surg. 2023;152(3):555E–556E. doi:10.1097/PRS.0000000000010545
67. Uzorka A, Namara S, Olaniyan AO. Modern technology adoption and professional development of lecturers. Educ Inf Technol. 2023;28(11):14693–14719. doi:10.1007/s10639-023-11790-w
68. DTK N, Leung JKL, Su J, RCW N, Chu SKW. Teachers’ AI digital competencies and twenty-first century skills in the post-pandemic world. Educ Technol Res Dev. 2023;71(1):137–161. doi:10.1007/s11423-023-10203-6
69. Casal-Otero L, Catala A, Fernández-Morante C, Taboada M, Cebreiro B, Barro S. AI literacy in K-12: a systematic literature review. Int J STEM Educ. 2023;10(1). doi:10.1186/s40594-023-00418-7
70. van de Vijver S, Tensen P, Asiki G, et al. Digital health for all: how digital health could reduce inequality and increase universal health coverage. Digit Heal. 2023;9(1):1–6. doi:10.1177/20552076231185434
link