Transforming Education: A Comprehensive Review of Generative Artiﬁcial Intelligence in Educational Settings through Bibliometric and Content Analysis

: In the ever-evolving era of technological advancements, generative artiﬁcial intelligence (GAI) emerges as a transformative force, revolutionizing education. This review paper, guided by the PRISMA framework, presents a comprehensive analysis of GAI in education, synthesizing key insights from a selection of 207 research papers to identify research gaps and future directions in the ﬁeld. This study begins with a content analysis that explores GAI’s transformative impact in speciﬁc educational domains, including medical education and engineering education. The versatile applications of GAI encompass assessment, personalized learning support, and intelligent tutoring systems. Ethical considerations, interdisciplinary collaboration, and responsible technology use are highlighted, emphasizing the need for transparent GAI models and addressing biases. Subsequently, a bibliometric analysis of GAI in education is conducted, examining prominent AI tools, research focus, geographic distribution, and interdisciplinary collaboration. ChatGPT emerges as a dominant GAI tool, and the analysis reveals signiﬁcant and exponential growth in GAI research in 2023. Moreover, this paper identiﬁes promising future research directions, such as GAI-enhanced curriculum design and longitudinal studies tracking its long-term impact on learning outcomes. These ﬁndings provide a comprehensive understanding of GAI’s potential in reshaping education and offer valuable insights to researchers, educators, and policymakers interested in the intersection of GAI and education.


Introduction
In an era where technological advancements are redefining human existence, there is one revolutionary force that stands out among the rest: artificial intelligence (AI).AI has permeated every corner of our lives, quietly and profoundly transforming the way we work, communicate, and navigate the world.The rise of AI was sparked in November 2022, when U.S. company OpenAI (San Francisco, CA, USA) released ChatGPT [1], an AI program that draws upon a large language database to generate responses from text-based inputs entered by humans [2].Since its launch, it has skyrocketed in popularity, becoming one of the fastest-growing consumer applications ever with an estimated 100 million active users every month [3].Following the successful release of ChatGPT, other tech companies have come out with their own AI-powered products, such as Google's Bard [4] and GitHub's Copilot [5].
AI has infused various domains of our lives, revolutionizing industries such as finance, education, engineering, healthcare, and many more.According to Hadi et al. [6], AI has transformed practices in the financial sector, such as algorithmic trading, market prediction, and financial reporting.Hadi et al. [6] further argue that education has seen the integration this review identifies research gaps and areas that require further investigation, guiding future research endeavors.The subsequent sections of this paper will, therefore, present the methodology employed in selecting and analyzing the research papers, followed by a comprehensive content analysis and synthesis of the findings and insights.Key themes, trends, and recommendations will be presented, allowing readers to gain a comprehensive understanding of the existing knowledge base.Subsequently, a comprehensive bibliometric analysis will be conducted, examining prominent AI tools, research focus, geographic distribution, and interdisciplinary collaboration in detail.Finally, this review will conclude by highlighting the implications of the synthesized findings and suggesting future research directions to advance the field of GAI in education.

Methodology
The methodology employed in this research follows a systematic approach to gather and analyze literature concerning the integration of GAI in education.It involves four key steps, as described below: a.
Literature retrieval-This step entails selecting appropriate search terms and keywords to comprehensively capture pertinent publications related to the intended topic representing the initial and pivotal phase of the data collection process.We retrieved a compilation of pre-existing articles and publications in the domain of GAI in education from the Scopus database.By utilizing a combination of the following keywords, "Education", "GAI", "Generative Artificial Intelligence", "GPT-4", "ChatGPT", "AlphaCode", "GitHub Copilot", and "Bard", the authors conducted a targeted search across titles, abstracts, and keyword fields.This process resulted in the enlistment of a total of 437 papers, spanning the years 2018 to 2023.b.
Literature screening-The literature screening process for this study drew inspiration from the PRISMA statement, which is widely recognized as a rigorous and transparent approach for conducting systematic reviews and meta-analyses (Figure 1).PRISMA provides a structured framework to ensure the systematic identification, selection, and evaluation of the relevant literature, promoting the reliability and reproducibility of the review process [15].Initially, a total of 437 papers were retrieved.Following the removal of duplicates, 312 papers remained.After observingly reviewing each publication, all articles that were deemed irrelevant to our research focus were discarded.We narrowed the list down to 217 reviews, papers, and publications that spanned from the years 2018 to 2023.The chronological growth of the papers (in the context of the topic) in the above-mentioned spanned years is depicted in Figure 2.
The figure above depicts a substantial number of papers available from our compiled list, specifically published in the year 2023.Notably, there is a significant and exponential growth in the field of GAI from 2018 to 2023, indicating a growing amount of research interest in the topic.This surge in research can be attributed to the recent emergence of popular GAI tools, such as ChatGPT, which have garnered considerable attention and sparked innovation within the education domain.The increasing research activity, therefore, suggests that GAI is perceived as a promising area for innovation in education and that researchers and educators are increasingly recognizing the potential benefits of integrating AI technologies into educational settings.

c.
Content analysis-This step entails carefully studying a large amount of information, like research publications, and organizing it in a meaningful way.This helps researchers see common themes and patterns in the information.In this case, the content analysis focused on how GAI is being used in education.The methodology involved categorizing and classifying the research papers into different themes and sub-themes.This organization helps researchers understand the different ways GAI is impacting education, making it easier to draw conclusions and insights from the collected information.
d. Bibliometric analysis-It serves the purpose of systematically evaluating academic literature, typically through analyzing citations and references in research papers.This methodology helps researchers understand the influence, trends, and relationships among different academic works.By examining patterns of citations, co-authorships, and keywords, bibliometric analysis allows for the identification of key authors, influential papers, emerging areas of study, and collaborative networks within a particular field.This approach aids researchers in gaining insights into the evolution of research topics, identifying leading scholars.
Sustainability 2023, 15, x FOR PEER REVIEW 4 of 40 sub-themes.This organization helps researchers understand the different ways GAI is impacting education, making it easier to draw conclusions and insights from the collected information.d.Bibliometric analysis-It serves the purpose of systematically evaluating academic literature, typically through analyzing citations and references in research papers.This methodology helps researchers understand the influence, trends, and relationships among different academic works.By examining patterns of citations, co-authorships, and keywords, bibliometric analysis allows for the identification of key authors, influential papers, emerging areas of study, and collaborative networks within a particular field.This approach aids researchers in gaining insights into the evolution of research topics, identifying leading scholars.The remaining part of this paper presents a content analysis and a bibliometric analysis while adhering to the methodological steps described in this section.

Content Analysis
The integration of GAI in education has become a subject of considerable interest, The remaining part of this paper presents a content analysis and a bibliometric analysis while adhering to the methodological steps described in this section.

Content Analysis
The integration of GAI in education has become a subject of considerable interest, opening up new possibilities and challenges for educational practices.GAI models, such as ChatGPT, have been extensively explored across various academic disciplines, presenting transformative opportunities in teaching, learning, and research.This section seeks to provide a thorough content analysis of the use of GAI in education.

Application of GAI in Computer Science
This study explores a collection of research papers focusing on the intersection of GAI and computer science education and assesses various aspects of GAI technologies and their implications for teaching and learning programming concepts (Table 1).
Table 1.The applications of GAI in computer science.

Themes
Authors Focus

Implications and potential of GAI code generation tools in programming education
Jonsson and Tholander (2022) [16] Highlights the transformative potential of generative machine learning in creative programming education, fostering both programming proficiency and creativity through students' exploration of the tool's inconsistent behavior.[17] Investigates the performance and implications of Codex, a deep learning model, in introductory programming, showcasing its ability to outperform students in problem-solving while raising considerations for computing education.[11] Explores the potential of large language models (LLMs) in providing explanations for programming concepts, contributing to the understanding of LLMs' role in enhancing learning experiences.Denny et al. (2023) [18] Investigates Copilot, an AI-powered code generation tool, emphasizing its problem-solving capabilities and discussing prompt engineering as a means to enhance computational thinking skills.Kazemitabaar et al. (2023) [19] Demonstrates the positive influence of AI code generators, like OpenAI Codex, on novice programmers, highlighting improvements in performance, computational thinking skills, and motivation.[20] Investigates the effectiveness of generative pre-trained transformer (GPT) models in answering code-related multiple-choice questions, highlighting the challenges faced by GPT models in code analysis and reasoning.Wermelinger (2023) [21] Focuses on GitHub Copilot, an AI-powered programming assistant, evaluating its performance and limitations in code generation, explanation provision, test generation, and bug fixing.

Ethical considerations and pedagogical approaches in AI education
Becker et al. ( 2023) [22] Emphasizes the need for collaborative efforts to explore the implications and effectiveness of AI-driven code generation tools in computing education, highlighting the importance of shaping their trajectory.
Jacques (2023) [23] Challenges the notion that AI-generated coding tools render human programming obsolete, advocating for a reevaluation of teaching methods that foster problem-solving, critical thinking, and communication skills.

Raji et al. (2021) [24]
Delves into the ethical considerations in AI education and proposes a shift toward a collaborative and holistic pedagogy that bridges the gap between computer science and humanities and social sciences.

Integration of AI technologies in programming education
Yilmaz and Yilmaz (2023) [25] Explores the incorporation of ChatGPT into programming education and highlights its positive impact on computational thinking skills, programming self-efficacy, and student motivation.
Cortinas-Lorenzo and Lacey (2023) [26] Addresses the need for transparency and interpretability in affective computing systems, highlighting the challenges associated with multimodal data, context integration, and interaction capturing.
The findings from the above publications reveal several key insights.Firstly, researchers highlight the transformative potential of generative machine learning in creative programming education, indicating that GAI can significantly impact how programming concepts are taught and learned.Secondly, crucial considerations are raised regarding AI-generated code, pointing to the need for careful examination of the quality and reliability of code produced by AI systems.Moreover, the potential of large language models (LLMs) in providing explanations for programming concepts is showcased, suggesting that AI can enhance the understanding of complex programming principles.Additionally, practical insights are offered into the implications of AI-powered code generation tools, like Copilot, shedding light on the benefits and challenges of using such technologies in educational settings.
These findings, therefore, provide valuable insights into GAI in computer science education, highlighting a lack of in-depth exploration of drawbacks and ethical concerns.While the spotlight is often on the positive outcomes of AI-generated code, it is crucial to acknowledge the possible drawbacks, like an excessive dependence on automated coding and its potential to undermine students' algorithmic comprehension.Furthermore, the studies focus on short-term improvements in performance and motivation without fully considering long-term effects on problem-solving skills and holistic learning.Future research should, therefore, balance the positive and negative aspects of GAI integration, addressing academic integrity, assessment challenges, and innovative methods for incorporating AI-generated code into programming curricula.

Application of GAI in Engineering Education
This section delves into a collection of research papers that explore the impact of GAI in various disciplines within engineering education.The papers are, therefore, categorized into three groups, as shown in Table 2.

Themes Authors Focus
Transformative role of AI in education and learning Marquez et al. (2023) [27] Emphasizes the potential of integrating product-based learning strategies and AI text generation models in biobased materials education, showcasing how this approach promotes active participation, problem-solving skills, and the generation of sustainable solutions.

Humphry and Fuller (2023) [28]
Demonstrates the utility of ChatGPT in revolutionizing chemistry education by effectively generating discussion sections for laboratory reports, highlighting its potential to improve student learning experiences and streamline the report writing process.

Sánchez-Ruiz et al. (2023) [29]
Investigates ChatGPT's impact on blended learning in engineering education, particularly mathematics.The study reveals that the integration of ChatGPT positively influences students' critical thinking abilities, problem-solving skills, and collaborative work.However, it also raises concerns about potential effects on the development of lateral competencies, highlighting the need for adaptive teaching strategies in blended learning environments.

Nikolic et al. (2023) [30]
Analyzes the impact of ChatGPT on assessment practices in engineering education.

AI in design and collaboration
Gmeiner et al. ( 2023) [31] Identifies challenges faced by designers when collaborating with AI-based design tools, including difficulties in understanding and adjusting AI-generated outputs and effectively communicating design objectives.
Chen et al. (2020) [32] Sheds light on the transformative influence of AI in engineering design education, enabling the realization of previously unattainable solutions and concepts.

Galdon et al. (2021) [33]
Presents an operational framework that harnesses abductive reasoning and probabilistic knowledge to enrich design and engineering education.The study showcases the framework's potential in fostering creative engagement with knowledge and generating impactful solutions, emphasizing the necessity of reframing probabilistic knowledge and embracing transformative approaches to address real-world uncertainties and challenges.

Innovative applications of social robots Elfaki et al. (2023) [34]
Introduces a cloud-based framework aimed at augmenting the intelligence and autonomy of social robots in engineering education.The study showcases the feasibility and potential of this framework, empowering social robots to showcase heightened capabilities through the utilization of cloud computing and clustering.
GAI presents transformative benefits in engineering education, as evident from various studies.Indeed, its integration enhances active learning and problem-solving skills and streamlines educational processes through advanced chatbots and text generation models.The potential to foster creative engagement in design and collaboration further empowers students to generate innovative solutions.Cloud-based frameworks amplify social robots' intelligence and accessibility in engineering education, providing new opportunities for interactive learning experiences.
Taking a critical view of these studies, while they undoubtedly showcase the positive implications of GAI for fostering active learning, honing problem-solving skills, and streamlining design processes, they also unveil crucial issues that demand careful consideration.These concerns encompass the potential erosion of lateral competencies due to overreliance on GAI, challenges in human-AI collaboration, and the necessity for a pedagogical framework that seamlessly incorporates GAI tools without diluting the essence of education.While these research papers accentuate the transformative potential of GAI in revolutionizing the educational landscape, they simultaneously underscore the urgency for judicious implementation, particularly within dynamic domains, like design and social robots.To address these issues, future work must refine GAI integration and assess long-term impacts to ensure its benefits are maximized while mitigating potential challenges.

Application of GAI in Higher Education
This section explores the potential, applications, challenges, and implications of GAI tools in specifically higher education.Table 3 shows the classification of these papers and their main focus.
Table 3.The application of GAI in higher education.

Publications Authors Focus
The integration of GAI tools such as ChatGPT in higher education

Konecki et al. (2023) [35]
Conducts a comprehensive examination of ChatGPT's role as an educational aid, providing insights into its benefits and limitations.
Eager and Brunton (2023) [36] Focuses on the broader implications of LLMs and conversational-style GAI on pedagogy.They emphasize the importance of implementing effective plagiarism detection mechanisms, fostering supportive learning environments, and offering faculty development programs to ensure the responsible integration of AI in teaching and learning.

Chaudhry et al. (2023) [37]
Intensively examines the impact of ChatGPT on learning outcomes and academic integrity in undergraduate degree programs.The study evaluates ChatGPT's performance in handling assignments and compares it to high-achieving students.Additionally, the effectiveness of plagiarism detection tools is assessed in the context of ChatGPT usage.The findings highlight the limitations of ChatGPT in its ability to fully replace human performance and underscore the need for a reevaluation of student evaluation approaches in light of AI-based tools.
Students' acceptance and use of GAI technology in higher education Iskender (2023) [38] Addresses the ethical and pedagogical considerations of GAI in higher education and conducted an interview with ChatGPT to delve into its impact on the realm of higher education and academic publishing.The interview sheds light on ChatGPT's potential in grading and supporting students, acknowledging its usefulness in certain educational contexts.However, it also raises crucial concerns about potential risks, such as the possibility of diminishing critical thinking skills and exacerbating educational inequalities.
Strzelecki (2023) [39] Delves into the adoption and use of ChatGPT among higher education students, developing a model based on technology adoption theory.The study identifies key predictors such as habit, performance expectancy, and hedonic motivation in determining students' acceptance and usage of ChatGPT.

Publications Authors Focus
Stojanov (2023) [40] Presents an autoethnographic study exploring the utilization of ChatGPT as a learning aid.The study underscores the potential benefits of ChatGPT while acknowledging its limitations, including occasional superficiality and inconsistency in generated answers.It emphasizes the need for caution and further research to fully understand the implications of incorporating ChatGPT in educational contexts.
The applications of AI and ML in higher education Pinto et al. (2023) [41] Conducts a systematic literature review to shed light on the diverse uses of ML in higher eduction.The review highlights the prediction of academic performance and employability as the most extensively studied application of ML in higher education.
The studies in this section offer insights into various aspects of GAI integration in higher education.They delve into the integration of GAI tools, such as ChatGPT and large language models, exploring their potential impact on teaching and learning practices.Student acceptance and usage of GAI technology are examined, revealing factors that influence their adoption.The applications of AI and machine learning in higher education are explored, particularly in predicting academic performance and employability.Furthermore, the impact of ChatGPT on learning outcomes and academic integrity is assessed, prompting discussions on assessment methodologies.Ethical and pedagogical considerations of GAI are discussed, highlighting both potential benefits and concerns related to originality and critical thinking.
These studies provide valuable insights into the potential of GAI tools in enhancing higher education.However, there are certain areas that could be more comprehensive, whereby the tendency to highlight the positive aspects of GAI integration could benefit from a more balanced exploration of potential challenges and unintended consequences.For instance, while the impact of tools like ChatGPT on student learning is explored, a deeper investigation into how such tools might affect critical thinking skills and originality in academic work could provide a more holistic understanding.Additionally, while the studies touch on students' acceptance of GAI technology, a deeper analysis of the sociocultural factors influencing students' attitudes and perceptions is warranted.By addressing these critical perspectives, future research in this area can provide a more nuanced and comprehensive understanding of the role and impact of GAI in higher education.

Application of GAI in Medical Education
This section aims to provide an overview of research papers that discuss the integration of GAI tools in healthcare, as shown in Table 4.
Table 4.The application of GAI in medical education.

Publications Authors Focus
The integration of GAI tools in healthcare Sallam et al. (2023) [9] Investigates the potential benefits and limitations of incorporating ChatGPT into medical fields.The study reveals the promising advantages of ChatGPT, such as personalized learning, clinical reasoning support, and better comprehension of complex concepts in medical education.
It highlights concerns related to data privacy, potential bias in generated content, and possible negative effects on critical thinking and communication skills.

Publications Authors Focus
Waisberg et al. (2023) [42] Investigates the potential applications of GPT-4 in neurosurgery, revealing its ability to enhance precision and effectiveness in neurosurgical practices, leading to improved patient outcomes and advancements in the field.
Sevgi et al. (2023) [43] Assesses the reliability of ChatGPT in neurosurgical education, pointing out that while it provides interesting responses, its credibility is compromised by the lack of citations for scientific queries, suggesting caution in relying solely on ChatGPT as an educational resource.
Šlapeta (2023) [44] Explores the advantages of integrating large language models, including ChatGPT, in parasitology education, highlighting their potential to improve learning outcomes and facilitate knowledge acquisition in the field.
Huh (2023) [45] Compares ChatGPT's knowledge and interpretive abilities with those of medical students in Korea in parasitology, revealing its inferior performance and indicating the need for further refinement in specialized medical contexts.

Dolezal et al. (2023) [46]
Investigates the use of deep neural networks and conditional generative adversarial networks (cGANs) to generate synthetic histology for tumor classification, showcasing its value as an educational tool for enhancing understanding of tumor biology.

Skalidis et al. (2023) [47]
Examines ChatGPT's performance in answering questions from a post-graduate exam in core cardiology, demonstrating its impressive success in handling higher-level examination content and raising discussions on its potential applications in diverse educational contexts within cardiology.

The applications of GAI in clinical education and practices
Corsello and Santangelo (2023) [48] Discusses the potential impact of ChatGPT on future pediatric research, highlighting positive effects such as improved clinical decision-making and enhanced medical education.However, the study also addresses concerns about bias, fairness, and ethical considerations associated with GAI utilization.

Lyu et al. (2023) [49]
Investigates the feasibility of utilizing ChatGPT for translating radiology reports into plain language, revealing its effective translation capabilities while noting occasional occurrences of random or oversimplified responses that require attention.

Abdel-Messih and Kamel Boulos (2023) [50]
Demonstrates ChatGPT's ability to successfully pass the United States Medical Licensing Examination (USMLE) by effectively answering queries related to clinical toxicology, showcasing its potential in this domain.

Oh et al. (2023) [51]
Explores the feasibility of integrating ChatGPT into surgical education and training, acknowledging its successful translation outcomes in radiology reports while also recognizing the need to address occasional occurrences of random or oversimplified responses.

Lee et al. (2023) [52]
Discusses the implications of GPT-4 for a chatbot specifically tailored for the field of medicine.The study examines the potential impacts of this well-educated chatbot on the practice of medicine.

Publications Authors Focus
The potential of GAI in health education and social interaction

Yang et al. (2023) [53]
Focuses on the utilization of ChatGPT in assisting children with Down syndrome, conducting a literature review to explore its advantages and limitations in various domains for individuals with this condition.It highlights the significance of ChatGPT as a valuable tool for promoting cognitive development, education, and social interaction, catering to the unique requirements of children with Down syndrome.

Wang et al. (2023) [54]
Explores the role of GAI in pandemic response and future medicine and conducts a comprehensive review of the potential applications of AI and GAI in the prevention and treatment of the COVID-19 pandemic.
The ethical considerations and responsible use of GAI in healthcare Cascella et al. (2023) [55] Investigates the potential applications and limitations of ChatGPT in healthcare, emphasizing the need to educate stakeholders about the appropriate utilization of AI-based language models in medicine while addressing the ethical challenges they may present.

Gupta et al. (2023) [56]
Examines the use of ChatGPT in generating innovative ideas in plastic surgery, highlighting its potential to contribute to research and support various aspects of patient care.However, ethical considerations must be taken into account to ensure its responsible and beneficial application.Masters (2023) [57] Addresses concerns about the improper utilization of ChatGPT by students and academics, emphasizing the importance of addressing this issue to prevent misuse and mitigate potential negative consequences.
The research papers in this section reveal a range of findings regarding the integration of GAI tools in medical education and practice.Notably, ChatGPT and GPT-4 showcase potential benefits, including personalized learning, improved comprehension of medical concepts, and effective translation of radiology reports.However, concerns about data privacy, potential bias, and impacts on critical thinking are brought to light.In clinical education, these tools exhibit promise in enhancing decision-making and supporting surgical training, although occasional random or oversimplified responses are identified as areas of caution.Specialized medical fields, such as neurosurgery and parasitology, show potential for GAI integration, but validation and context-specific refinement are crucial.The application of GAI in health education for individuals with Down syndrome highlights its significance in cognitive development and social interaction.Moreover, GAI's role in pandemic response and shaping future medicine proves noteworthy, with AI models contributing to mitigating COVID-19 spread and suggesting new research directions.Ethical considerations emerge as a common theme, emphasizing responsible usage and preventing misuse of GAI tools in healthcare.
A critical perspective reveals that these papers sometimes tend to overlook significant concerns such as data privacy, biases, and the risk of undermining critical thinking.Consequently, future work should prioritize establishing ethical guidelines and governance frameworks to address data privacy, bias, and transparency.Rigorous validation processes are essential to ensure the accuracy and reliability of GAI-generated information.Continuous research and development efforts are needed to enhance GAI capabilities for more complex medical scenarios.Interdisciplinary collaboration can promote responsible and effective integration of GAI in medical practices and education.

Application of GAI in Nursing Education
In this section, we explore a few research papers that delve into the implications of ChatGPT in nursing education (Table 5).
Table 5.The application of GAI in nursing education.

Publications
Authors Focus

Potential and implications of ChatGPT
Choi et al.'s (2023) [58] Examines the effects of ChatGPT on nurse education, highlighting the importance of striking a balance between AI integration and human interaction to ensure comprehensive learning experiences.

Irwin et al. (2023) [59]
Discusses the broader implications of ChatGPT in nursing and midwifery practice, stressing the potential benefits it brings alongside the challenges in implementing this transformative technology.

Ethical considerations of using ChatGPT in nursing education
Abdulai and Hung (2023) [60] Delves into the ethical considerations of using ChatGPT in nursing education, emphasizing the significance of aligning AI integration with nursing's core values to maintain the profession's unique identity.
Fostering critical thinking skills among nursing students.

Seney et al. (2023) [61]
Sheds light on ChatGPT's role in enhancing clinical judgment in nursing education and proposes a thoughtful teaching strategy to address its limitations in generating NCLEX-style questions.
Ahmed (2023) [62] Takes a closer look at ChatGPT's transformative impact on personalized learning and clinical decision-making for nurses while addressing the ethical considerations surrounding its use.

Lo (2023) [63]
Introduces the CLEAR Framework, a valuable tool for equipping nursing students with critical thinking skills, enabling them to engage effectively with AI language models and promoting responsible AI use in nursing education.
The research papers in this section focus on the potential benefits of integrating GAI into nursing education.These studies explore how AI can enhance patient care, information literacy, and critical thinking skills among nursing students.The papers emphasize the importance of balancing AI integration with human interaction for comprehensive learning experiences and address ethical considerations.They delve into topics such as AI's impact on clinical judgment, broader implications in nursing practice, personalized learning, and clinical decision-making.Frameworks are introduced to equip nursing students with critical thinking skills for responsible AI engagement.
Future work in this domain is needed to address critical issues such as data privacy, biases, and the potential threat of compromising human-centered nursing practices.This can be achieved by establishing robust ethical guidelines and governance frameworks for the integration of AI in nursing education.Additionally, rigorous validation processes are essential to ensure the accuracy and reliability of AI-generated content.

Application of GAI in Communication
In this section, we explore a collection of research papers that delve into the applications and implications of GAI in communication education.These studies shed light on how GAI models are revolutionizing research, education, journalism, and social media in the context of communication, as shown in Table 6.Table 6.The application of GAI in communication.

Publications Authors Focus
GAI models are revolutionizing in the context of communication.
Wang (2022) [64] Focuses on machine translation, specifically examining the use of generative adversarial network-based neural machine translation models to improve Chinese-English translation accuracy and naturalness.
Santandreu-Calonge et al. ( 2023) [65] Investigates the advantages of natural language processing tools, like ChatGPT, in addressing challenges such as jargon and coordination issues.The authors emphasize the importance of combining AI technologies with human expertise to achieve effective communication and enhanced patient outcomes.
The analysis highlights the transformative capacity of GAI in shaping the future of journalism and media, highlighting both the opportunities and challenges that arise from the integration of AI in these fields.
Their research uncovers the profound influence of ChatGPT on streaming media and highlights the potential of integrating ChatGPT with visual generative models.
These studies showcase the effectiveness of GAI in improving machine translation accuracy and its transformative potential in journalism and media education.Furthermore, the impact of AI on streaming media is explored, providing an understanding of GAI's role in communication education.
The small number of papers provide a restricted view of the challenges and ethical issues linked to biases and human interaction.These papers predominantly accentuate the positive impacts of GAI.This highlights the need for further research to gain a holistic understanding of GAI's impacts on communication and strike a balance between technological advancements and maintaining authentic human engagement.Future research could focus on developing and assessing AI-powered tools for communication skills, as well as examining ethical implications and potential biases in GAI model implementation in communication and journalism.

Application of GAI in Academia
This section examines a collection of research papers focusing on GAI's impact on research and its potential applications in different fields, as summarized in Table 7 below.
This section highlights the transformative impact of GAI in academia, showcasing its potential benefits for higher education, research, and library services.The discussion covers the advantages of AI technologies, natural language processing techniques, and large language models in various domains.However, it also stresses the need to address ethical concerns, algorithmic biases, and maintain academic integrity.Strategies to ensure integrity and accuracy are explored to address these challenges.Additionally, the section sheds light on the implications of GAI in library science.
Moving forward, future research should take a critical and interdisciplinary approach to GAI integration in academia, considering a wide array of potential benefits and risks.It is imperative to delve deeper into the ethical implications, potential biases, and algorithmic transparency associated with GAI tools.Moreover, a more in-depth investigation into the long-term effects of GAI on academic integrity, learning experiences, and research methodologies is warranted.

Publications Authors Focus
GAI and ethical considerations in academia Alqahtani et al. (2023) [68] Highlights the transformative impact of (GAI) models, such as GPT-4 and BARD, in higher education and research.They explore the potential benefits of these technologies, including text generation, data analysis, literature review, and educational support.However, the authors also emphasize the importance of addressing ethical concerns and algorithmic biases to ensure responsible and effective utilization of GAI in academia.

Cotton et al. (2023) [69]
Sheds light on the benefits and challenges associated with the use of ChatGPT in higher education.The authors emphasize the advantages of increased student engagement and accessibility while addressing concerns related to academic dishonesty.Strategies to detect and prevent cheating are explored, highlighting the need for a proactive and ethical approach to integrating AI tools.Currie (2023) [70] Focuses on the impact of ChatGPT on academic integrity in nuclear medicine and radiology, acknowledging its potential while highlighting limitations, such as errors and fabrications.Emphasizes the ethical and responsible integration of ChatGPT.

Dergaa et al. (2023) [71]
Examines the prospects and potential threats of ChatGPT in academic writing, emphasizing the need to redefine norms and adjust information expectations for ethical utilization.Perkins (2023) [72] Explores the implications of AI tools like ChatGPT on academic integrity during formal assessments, addressing concerns about undetectable original content generation.The importance of transparent student use and updates to academic integrity policies in higher education institutions are emphasized throughout these papers.

GAI in library science
Inamdar (2023) [73] Conducts a comprehensive investigation into the impact of artificial intelligence text generators (AITGs) on library services, resources, and staff roles, highlighting their potential to enhance various aspects of library operations and accessibility.
Frederick (2023) [74] Focuses on information environments and explores the role of ChatGPT in libraries, analyzing current discussions and historical developments.The paper uncovers the potential risks and benefits associated with ChatGPT, providing valuable guidance for librarians to navigate the implications of this technology and emphasizing the importance of librarians acquiring knowledge about AI to adapt to changing information landscapes.

Application of GAI in General Education
In this section, we delve into the general applications of GAI in education.While earlier sections have explored the use of GAI in specific domains, such as medical education, higher education, engineering education, and more, this section broadens the scope to encompass the wider applications of GAI in education.To provide a structured analysis of this topic, the content is divided into several subsections.This section, therefore, covers a review of a wide range of applications and considerations regarding the integration of GAI in education, as discussed in Table 8 below:  [117].

•
The need for adaptation, heightened awareness, and thoughtful ethical considerations to harness AI's positive impact.

•
A pedagogical approach to equip learners with skills to navigate AI systems.

•
Understanding societal boundaries and ethical standards to foster critical engagement.

•
Opportunities, threats, and limitations of ChatGPT, and the importance of human expertise.

•
Balance between AI-assisted innovation and human expertise.

•
The need for multidisciplinary research to understand and harness its potential.

•
Positive applications and ethical implications of AI-generated characters, including design considerations and potential impact in entertainment, education, and therapy.

•
Understanding GAI techniques and assessing the societal implications of misinformation through a curriculum for middle school students.

•
Human-like avatars, gamification elements, and emotional intelligence in chatbot design to further enhance learning outcomes.

•
Enhanced learning outcomes through the integration of ChatGPT into education.

•
ChatGPT for character development and authentic assessments in higher education.

•
Senior computer engineering students' perceptions of using ChatGPT in learning activities.

•
The importance of communication and teamwork skills is highlighted by the value of oral exams in physical chemistry classrooms.

•
The need for improved communication and information exchange within the classroom.

•
Influence of factors such as students' perception, satisfaction, and engagement on the utilization of ChatGPT for personalized tutoring.

•
Utilization of ChatGPT and social media for collecting real-time student feedback on teaching styles in higher education.

•
Collaborative nature of teacher-AI interaction and the importance of teachers' pedagogical expertise.

•
Influence of learning from an AI-generated virtual instructor resembling an admired individual.• Vulnerability of STEM learners to AI-generated misinformation.

•
Perspectives of students and educators on using AI.

•
Analyzed the concerns expressed by educators regarding the integration of ChatGPT in educational settings.

•
Proposed the BAG model (Build, Assess, and Govern) as a teaching paradigm.

•
A comprehensive framework to enhance peer feedback processes.

•
Challenges of teaching quantitative literacy and reasoning alongside language-based AI systems, like ChatGPT.

•
Teachers' adaptability and creativity in incorporating AI tools into their instructional approaches.

•
Gaps in curriculum and educational policies, emphasizing the need for interdisciplinary approaches.

•
Students' preferences regarding university teaching methods.
The remaining part of this section synthesizes and discusses the applications of GAI in general education across each subcategory of papers outlined in Table 8.
Technology Adoption and Integration in Education-The findings demonstrate improved learning outcomes, enhanced student engagement, and the potential for personalized learning experiences.However, ethical considerations, responsible usage, and addressing limitations are crucial.Future work necessitates a more critical perspective in developing nuanced guidelines and policies, cautiously navigating interdisciplinary collaboration and fostering genuinely informed critical thinking skills.This is imperative to ensure that AI integration in education addresses complex challenges effectively and ethically.Additionally, an in-depth exploration of the lasting repercussions of AI technologies on education is crucial, surpassing superficial assessments and delving into their potential to reshape fundamental educational dynamics.
Classroom Design and Learning Environment-The papers in this category highlight the significance of incorporating technology and GAI in creating effective learning environments.The research demonstrates the potential of AI models to automate classroom design, optimize learning spaces, and enhance the learning experience through interactive dialogue systems and gamified environments.However, it is crucial to acknowledge the potential pitfalls, like algorithmic biases and copyright concerns.Moving forward, there is a pressing need to prioritize the promotion of AI interactions that are inclusive and culturally sensitive.Moreover, the development of AI-powered solutions tailored to resource-limited educational contexts should be a key focus of future endeavors.
Assessment Practices and Academic Integrity-The research highlights the superior performance of AI chatbots, such as ChatGPT, in subject-specific exams and the potential of natural language processing techniques to enhance peer feedback processes.However, concerns regarding authenticity, credibility, systematic errors, and the risk of plagiarism are raised, underscoring the need for caution and the responsible integration of AI tools in educational assessments.While the papers touch on the importance of updating assessment strategies, policy implementation, instructor training, and student awareness about AI's impact on evaluations, a more in-depth analysis of these measures' efficacy is warranted.To advance, it is crucial to fine-tune AI models for assessment precision, tackle biases and accuracy concerns, and establish robust guidelines and ethical standards to govern AI's role in educational assessments.
Language Instruction-The publications demonstrate how integrating AI chatbots into language instruction can offer personalized support, optimize learning experiences, and address diverse language proficiencies.However, a more critical viewpoint uncovers the apprehensions surrounding the preservation of academic integrity and equitable learning experiences.While the potential gains are evident, the unexplored territory of AI's potential biases and fairness concerns necessitates not only policy frameworks but also a deep reevaluation of instructional strategies.
Ethics and Responsible GAI Use-These papers highlight the importance of ethical considerations and the responsible use of AI systems, particularly chatbots like ChatGPT, in education and research.These works shed light on the merits, shortcomings, obstacles, and debates surrounding such technologies.Looking ahead, it is imperative that forthcoming efforts concentrate on interdisciplinary investigations to confront ethical difficulties, integrate conscientious AI education into academic curricula, and promote collaborative endeavors aimed at formulating comprehensive regulations and principled directives.
Subject-Specific Education and Skills Development-The reviewed research papers shed light on the potential applications of GAI in specific educational domains, including sustainability education, digital skills development, chemistry education, biology and environmental science, and construction education.The findings indicate that GAI technologies, such as ChatGPT, have the ability to simplify complex tasks, expedite education and research processes, enhance hazard recognition capabilities, promote safety management practices, and facilitate interdisciplinary approaches.However, the papers also acknowledge the limitations and potential risks associated with the utilization of GAI, such as concerns about data privacy, biases, and the need for human engagement.Moving forward, future work should focus on the seamless integration of sustainability education into AI learning frameworks, a dedicated commitment to addressing issues of equity and accessibility to ensure that the benefits of GAI are universally accessible, and a diligent pursuit of strategies to maximize the advantages that GAI offers while concurrently taking proactive steps to mitigate the inherent risks that come with its implementation.
Learning Outcomes and Student Engagement-The papers highlight the positive impact of GAI, such as AI chatbots and virtual instructors, on students' learning outcomes and engagement.The findings demonstrate improvements in learning outcomes, motivation, and engagement when utilizing AI chatbots, like ChatGPT.However, a more discerning analysis beckons for heightened scrutiny in various aspects.This includes the imperative refinement of AI chatbot design, probing into alternative assessment strategies, enhancing the interaction between AI and educators, and undertaking an in-depth exploration into the long-term impact of AI on student motivation and engagement.
Challenges and Considerations in AI Integration-The related papers shed light on the obstacles and considerations associated with integrating GAI, such as ChatGPT, into education.These studies emphasize the importance of addressing concerns related to AI-generated misinformation, ensuring the secure integration of AI in education, and promoting responsible implementation practices.The findings underscore the need for raising learners' awareness about the potential influence of AI on their learning and further research to ensure the conscientious and secure integration of AI technologies, like ChatGPT, in educational settings.The future trajectory in this field involves developing comprehensive strategies and guidelines to address challenges while conducting further research on AI integration's impact on student learning outcomes.This avenue of exploration aims to formulate effective approaches to counter challenges and rigorously investigate how AI integration influences students' academic achievements.
Pedagogical Approaches and Teaching Methods-The papers shed light on innovative instructional practices and frameworks that integrate AI technologies into education.The findings offer insights into teaching ethics in AI courses, enhancing peer feedback processes through natural language processing techniques, addressing challenges in teaching quantitative literacy alongside language-based AI systems, and showcasing the integration of AI to support student learning.The papers also emphasize the importance of interdisciplinary collaboration.Future endeavors could dedicate their focus to meticulously refining and substantially expanding the currently presented pedagogical methodologies.Delving deeper, a crucial area of investigation could encompass understanding the enduring effects of AI integration on the overall learning achievements of students, shedding light on potential shifts in learning dynamics, cognitive development, and critical thinking skills over the long term.Furthermore, it becomes imperative to navigate the intricate field of ethical considerations, delving into the multifaceted implications of AI technologies within educational contexts.
To sum-up the findings in this section, Figure 3 provides a comprehensive visualization and categorization of the diverse applications of GAI in the field of education.This spider graph employs a two-tier structure to present the information effectively.The initial tier highlights the primary themes that have been central to a majority of applications.The proportional sizing of the circles reflects the prevalence of these themes, representing the number of applications and associated publications in each domain.Furthermore, within each thematic category, a secondary layer delineates specific keywords, encapsulating the nuances of each area.The size of these sub-circles corresponds to the prominence of these keywords, which are characterized by their density and the number of related research publications.This elaborate spider graph succinctly encapsulates the breadth and depth of GAI's influence in education, providing an illuminating snapshot of the ongoing research endeavors and practical implementations in this field.

Bibliometric Analysis
In this section, we embark on a comprehensive bibliometric analysis that delves into the field of GAI within the context of education.By examining a vast collection of existing articles and publications retrieved from the Scopus database, we seek to provide a deeper understanding of the current state of research and innovation in this evolving field.Through systematic literature screening and data analysis, we aim to uncover valuable insights, trends, and implications surrounding GAI's integration and impact on teaching and learning in educational settings.This exploration will shed light, for example, on the interdisciplinary nature of GAI research, key AI tools employed, the geographic distribution of contributions, and the diverse fields within education where GAI finds its applications.As we navigate through the findings and visualizations, we uncover the growth trajectory of GAI research, observe patterns of co-occurrence between terms, and ascertain the prominent applications of GAI tools, like ChatGPT, in educational settings.
To provide an in-depth analysis of the current state of GAI in education, we utilized VOSviewer to generate five types of visualization maps.These maps are designed to

Bibliometric Analysis
In this section, we embark on a comprehensive bibliometric analysis that delves into the field of GAI within the context of education.By examining a vast collection of existing articles and publications retrieved from the Scopus database, we seek to provide a deeper understanding of the current state of research and innovation in this evolving field.Through systematic literature screening and data analysis, we aim to uncover valuable insights, trends, and implications surrounding GAI's integration and impact on teaching and learning in educational settings.This exploration will shed light, for example, on the interdisciplinary nature of GAI research, key AI tools employed, the geographic distribution of contributions, and the diverse fields within education where GAI finds its applications.As we navigate through the findings and visualizations, we uncover the growth trajectory of GAI research, observe patterns of co-occurrence between terms, and ascertain the prominent applications of GAI tools, like ChatGPT, in educational settings.
To provide an in-depth analysis of the current state of GAI in education, we utilized VOSviewer to generate five types of visualization maps.These maps are designed to provide a comprehensive analysis of the topic.Each map consists of circles that represent different items such as publications, researchers, or terms.The size of the circle and the font within it indicate the level of activity associated with that item.A larger circle and a bigger font size indicate a higher level of activity, whereas a smaller circle and a smaller font size suggest less activity.The distance between any two terms in the diagram reflects the degree of association between them.A shorter distance signifies a stronger correlation, whereas a longer distance indicates a weaker correlation [151].

Co-Occurrence Map Based on Text Data
The relevant and frequently occurring terms were identified by analyzing the text data of the 217 selected publications.The analysis of text data looks to extract relevant terms from the titles and abstracts of the selected articles and constructs a network of co-occurrence links between them [151].This analysis enables us to detect any emerging developments and identify impactful terms within the field of GAI in education.A total of 4408 terms were generated, out of which 102 terms meet the minimum threshold of 10 occurrences.To further narrow down the selection of terms, VOSviewer calculated a relevance score for each term, and the top 60% most relevant terms were selected, resulting in the 61 terms that were displayed in the network seen in Figure 4. Terms with a high relevance score represent more specific topics covered by the text data, while terms with a low relevance score tend to be of a general nature [151].A thesaurus was created and uploaded to VOSviewer for the standardization of keywords and the elimination of duplicate terms [152].while "Question" signifies GAI's potential to create queries and foster interactive learning environments."Concern" highlights the importance of addressing ethical and practical considerations when employing GAI in education."Opportunity" underscores the positive prospects that GAI offers in enhancing educational experiences and outcomes.The prominence of "Generative Pre-training Transformers (GPT)" demonstrates the pivotal role of advanced language models, like ChatGPT, in driving GAI applications.Terms like "Assessment" and "Capability" indicate GAI's relevance in evaluating student performance and its potential to augment educational capabilities.This analysis unveils the nuanced facets of GAI in education, encompassing both its capabilities and the essential aspects that demand attention for responsible implementation.The findings in Figure 4 demonstrate the diverse range of research and exploration within the field of GAI in education.The interconnected web of key terms suggests that the research on GAI has encompassed the areas of assessment, teaching, research, higher education, academic integrity, and ethics.The findings also highlight GAI's revolutionary potential with the appearance of terms such as "opportunity" and "effectiveness".Additionally, the terms that are directly linked with GAI are shown in Figure 5.The findings in Figure 5 indicate a strong relationship between GAI and "Generative Pre-training Transformers" (GPT), which are a type of large language model (LLM).GPTs, such as ChatGPT, have emerged as foundational technologies in GAI and have demonstrated their capabilities in various applications.The integration of GPTs in educational contexts, particularly in the development of AI chatbots, has the potential to enhance learning experiences through personalized interactions, intelligent feedback, and adaptive learning pathways.This raises questions about the impact of GAI on student engagement, motivation, and learning outcomes.Additionally, the association between GAI and terms such as opportunity, question, response, and assessment highlight the diverse applications of GAI in educational contexts.GAI has the potential to support various aspects of teaching and learning, including generating questions, providing responses, and assessing student performance.The use of GAI in assessments raises concerns about the fairness and reliability of automated grading systems.It is crucial to examine the validity and equity of GAI-based assessment methods to ensure that they align with established educational standards and principles of fairness.
The ten most frequently occurring terms and their number of occurrences can be seen listed below in Table 9.Also listed in the table is the relevance score for each term.The relevance score for each term is calculated by VOSviewer.For each term, VOSviewer determines the distribution of (second order) co-occurrences over all terms and then compares it to the overall distribution of co-occurrences over terms.The larger the difference between the two distributions (measured using the Kullback-Leibler distance), the higher the relevance of a term [153].Table 9 provides an insightful analysis of the key terms frequently encountered in the context of GAI in education.These terms shed light on the diverse aspects of GAI's integration into educational practices.The term "System" suggests the broader framework within which GAI operates, emphasizing the need for systematic implementations."Response" indicates the generation of AI-driven outputs, while "Question" signifies GAI's potential to create queries and foster interactive learning environments."Concern" highlights the importance of addressing ethical and practical considerations when employing GAI in education."Opportunity" underscores the positive prospects that GAI offers in enhancing educational experiences and outcomes.The prominence of "Generative Pre-training Transformers (GPT)" demonstrates the pivotal role of advanced language models, like ChatGPT, in driving GAI applications.Terms like "Assessment" and "Capability" indicate GAI's relevance in evaluating student performance and its potential to augment educational capabilities.This analysis unveils the nuanced facets of GAI in education, encompassing both its capabilities and the essential aspects that demand attention for responsible implementation.

Co-Occurrences Map Based on Keywords
The keywords that occur frequently were identified by analyzing the bibliographic data of the 217 selected publications.A total of 1204 keywords were found, out of which 50 were selected, where the minimum number of occurrences of those terms was limited to 6 (Figure 6).A thesaurus was also used in VOSviewer to standardize keywords and eliminate redundant duplicate terms.This analysis looks at all keywords, including both author keywords and index keywords.Author keywords are specific keywords chosen and provided by the authors of the articles themselves, while index keywords are keywords that have been assigned by indexers or databases to classify and categorize the articles for the purpose of information retrieval and indexing.Some of the most common keywords identified include "Artificial intelligence", "Chat-GPT", "Education", and "Human".A list of the ten most frequently used keywords, their occurrences, and their total link strength can be found in Table 10.The total link strength indicates the total strength of the co-occurrence links of a given keyword with other keywords [151].
Table 10 presents a comprehensive analysis of the most frequently occurring keywords in the domain of GAI in education, offering valuable insights into the prominent themes and research areas.The top ten keywords are indicative of the central focus and key aspects of GAI's integration into educational contexts."Artificial intelligence" emerges as the most prevalent keyword, reflecting the overarching role of AI technologies in transforming educational practices."ChatGPT" follows closely, showcasing the widespread recognition and utilization of this AI chatbot in educational applications."Education" and "Human" hold significant positions, highlighting the significance of GAI in shaping learning experiences and considering human aspects in the educational process."Education computing" signifies the intersection of computing technologies and education, while "Machine learning" underscores the pivotal role of ML algorithms in GAI applications."Medical education" and "Natural language processing" reveal GAI's potential in healthcare education and language-based learning environments.Lastly, "Students" emphasizes the focus on learners and their experiences, reflecting the student-centric approach that GAI-driven educational interventions strive to achieve.These keywords depict the breadth and depth of GAI research in education, with an emphasis on AI-driven technologies, human-centric design, and transformative potential in diverse educational settings.

Co-Occurrences Map Based on Keywords
The keywords that occur frequently were identified by analyzing the bibliographic data of the 217 selected publications.A total of 1204 keywords were found, out of which 50 were selected, where the minimum number of occurrences of those terms was limited to 6 (Figure 6).A thesaurus was also used in VOSviewer to standardize keywords and eliminate redundant duplicate terms.This analysis looks at all keywords, including both author keywords and index keywords.Author keywords are specific keywords chosen and provided by the authors of the articles themselves, while index keywords are keywords that have been assigned by indexers or databases to classify and categorize the articles for the purpose of information retrieval and indexing.Some of the most common keywords identified include "Artificial intelligence", "ChatGPT", "Education", and "Human".A list of the ten most frequently used keywords, their occurrences, and their total link strength can be found in Table 10.The total link strength indicates the total strength of the co-occurrence links of a given keyword with other keywords [151].To further the analysis, two additional keyword maps were generated.Firstly, a map of the most common author keywords was generated.The keywords were limited to a minimum of five occurrences, which resulted in a total of twenty-five author keywords in the network visualization (Figure 7).Similarly, a map of the most common index keywords was created.The keywords were also limited to a minimum of five occurrences, which resulted in a total of thirty-nine index keywords in the map (Figure 8).
To further the analysis, two additional keyword maps were generated.Firstly, a map of the most common author keywords was generated.The keywords were limited to a minimum of five occurrences, which resulted in a total of twenty-five author keywords in the network visualization (Figure 7).Similarly, a map of the most common index keywords was created.The keywords were also limited to a minimum of five occurrences, which resulted in a total of thirty-nine index keywords in the map (Figure 8).The analysis of keywords reveals the central focus of GAI research in education, with terms such as artificial intelligence, ChatGPT, and education featuring prominently.This emphasizes the significance of AI technologies in transforming educational practices and the need for collaborations between AI experts and educators.The inclusion of medical education-related keywords suggests potential applications of GAI in healthcare education.GAI has the potential to support medical training, enhance diagnostic processes, and improve patient care.

Co-Occurrence Map Based on Country of Co-Authorship
Furthermore, the data analysis enabled us to examine the geographic distribution of these publications.The visualization map was generated for the countries of co-authorship by setting the minimum number of documents of a country to five.Of the 69 countries with publications, 20 met the threshold, and the results are depicted in Figure 9.The analysis of keywords reveals the central focus of GAI research in education, with terms such as artificial intelligence, ChatGPT, and education featuring prominently.This emphasizes the significance of AI technologies in transforming educational practices and the need for collaborations between AI experts and educators.The inclusion of medical education-related keywords suggests potential applications of GAI in healthcare education.GAI has the potential to support medical training, enhance diagnostic processes, and improve patient care.

Co-Occurrence Map Based on Country of Co-Authorship
Furthermore, the data analysis enabled us to examine the geographic distribution of these publications.The visualization map was generated for the countries of co-authorship by setting the minimum number of documents of a country to five.Of the 69 countries with publications, 20 met the threshold, and the results are depicted in Figure 9.

Co-Occurrence Map Based on Country of Co-Authorship
Furthermore, the data analysis enabled us to examine the geographic distribution of these publications.The visualization map was generated for the countries of co-authorship by setting the minimum number of documents of a country to five.Of the 69 countries with publications, 20 met the threshold, and the results are depicted in Figure 9.Most publications appeared from English-speaking countries (USA, Australia, and the UK), followed by India, Germany, and China.The ten countries with the strongest links on the map are listed in Table 11.Table 11 presents the top 10 countries based on their contributions to GAI research in education.The United States leads the list, indicating its central role in GAI-related studies.Australia and the United Kingdom follow closely, showcasing their active engagement in the field.New Zealand, India, Germany, Hong Kong, Italy, Saudi Arabia, and Switzerland also feature, reflecting their emerging roles in GAI research within the educational context.
The geographical distribution of publications demonstrates widespread global interest and collaboration in exploring the potential of GAI in education.Moreover, the distribution indicates the dominance of English-speaking countries such as the United States, Australia, and the United Kingdom.This aligns with the global prominence of these countries in educational research and technological advancements.However, it is worth noting the increasing contributions from countries like India, Germany, and China, suggesting a broader global interest in GAI in education.Collaboration and knowledge sharing across countries and regions can further propel advancements and foster cross-cultural perspectives in GAI research.

Co-Occurrence Map Based on Authorship
A further analysis looks at investigating the authorship relations across the multitude of authors of the publications.The network visualization map was limited to authors with a minimum of five citations.Out of the 682 authors, 144 met the threshold, and the largest set of connected authors included 23 authors.The resulting network visualization, depicted in Figure 10, presents a complex web of links connecting these 23 authors.
The top 10 authors in terms of documents produced and citations can be seen in Tables 12 and 13, respectively.The total link strength in each table indicates the total strength of the co-occurrence links of a given author with other authors [151].The network visualization consists of a large network of links connecting the authors, which showcases the collaborative and interdisciplinary nature of the research being conducted within the field of GAI in education.Moreover, it demonstrates how researchers from diverse backgrounds and institutions come together to advance knowledge and innovation in this area.By highlighting these connections, the visualization fosters a better understanding of the knowledge-sharing dynamics and the collective effort behind the advancements in GAI in education.This insight can prove invaluable in identifying key contributors, potential research trends, and opportunities for further collaboration and development within the field.
The top 10 authors in terms of documents produced and citations can be seen in Tables 12 and 13, respectively.The total link strength in each table indicates the total strength of the co-occurrence links of a given author with other authors [151].
Table 12 depicts the key contributors of research in the domain of GAI in education.The relatively high number of documents, citations, and total link strength demonstrates the great influence and contribution of the authors.Similarly, Table 13 identifies key contributors who have been influential through their publications within this domain.Notably, the top four authors, namely Hwang G.-J., Chen X., Xie H. and Zou D., all have an impressive citation count of 162.This exceptional citation count comes from their highly influential research paper titled "Application and theory gaps during the rise of Artificial Intelligence in Education", which played a vital role in exploring the applications of AI in education in 2020, contributing significantly to the early development of the field [89].

Data Analysis on Article Sources
To investigate the different sources of the 217 publications, a list of all unique sources was created and then ordered from the highest number of articles to the lowest.A total of 152 unique sources were found.The sources with the highest number of publications were "Computers and Education: Artificial Intelligence" and the "Journal of University Teaching and Learning Practice", with five publications each.A bar graph showcasing the top 10 sources by number of publications can be found in Figure 11.Table 12 depicts the key contributors of research in the domain of GAI in education.The relatively high number of documents, citations, and total link strength demonstrates the great influence and contribution of the authors.
Similarly, Table 13 identifies key contributors who have been influential through their publications within this domain.Notably, the top four authors, namely Hwang G.-J., Chen X., Xie H., and Zou D., all have an impressive citation count of 162.This exceptional citation count comes from their highly influential research paper titled "Application and theory gaps during the rise of Artificial Intelligence in Education", which played a vital role in exploring the applications of AI in education in 2020, contributing significantly to the early development of the field [89].

Data Analysis on Article Sources
To investigate the different sources of the 217 publications, a list of all unique sources was created and then ordered from the highest number of articles to the lowest.A total of 152 unique sources were found.The sources with the highest number of publications were "Computers and Education: Artificial Intelligence" and the "Journal of University Teaching and Learning Practice", with five publications each.A bar graph showcasing the top 10 sources by number of publications can be found in Figure 11.The analysis of article sources, therefore, presents a compelling picture of the interdisciplinary nature of GAI research in education, drawing insights from various fields such as computer science, education, and even the medical field.The convergence of expertise from diverse disciplines underscores the significance of collaborative efforts in addressing the complex challenges associated with GAI implementation in education.

Data Analysis on Document Type, Fields, GAI Tools Used, and Research Types
In this section, we classified the 217 publications based on their type, field of study, GAI tools used, and paper types.Figure 12 presents an analysis of the occurrences of various document types from Scopus.Notably, the most common document type is "Article", The analysis of article sources, therefore, presents a compelling picture of the interdisciplinary nature of GAI research in education, drawing insights from various fields such as computer science, education, and even the medical field.The convergence of expertise from diverse disciplines underscores the significance of collaborative efforts in addressing the complex challenges associated with GAI implementation in education.

Data Analysis on Document Type, Fields, GAI Tools Used, and Research Types
In this section, we classified the 217 publications based on their type, field of study, GAI tools used, and paper types.Figure 12 presents an analysis of the occurrences of various document types from Scopus.Notably, the most common document type is "Article", which appears 99 times, highlighting its preeminent role as the primary form of publication within the dataset."Article" is followed by "Note", with 23 occurrences, and "Conference Paper" and "Review", with 20 and 18 occurrences, respectively.Moderately represented are "Letter" and "Editorial", each occurring 17 and 11 times, respectively.Conversely, "Short Survey" is the least common, appearing only once.The dominance of scholarly articles in the dataset aligns with the typical composition of academic databases like Scopus, emphasizing their pivotal role as reliable sources of scientific information.which appears 99 times, highlighting its preeminent role as the primary form of publication within the dataset."Article" is followed by "Note", with 23 occurrences, and "Conference Paper" and "Review", with 20 and 18 occurrences, respectively.Moderately represented are "Letter" and "Editorial", each occurring 17 and 11 times, respectively.Conversely, "Short Survey" is the least common, appearing only once.The dominance of scholarly articles in the dataset aligns with the typical composition of academic databases like Scopus, emphasizing their pivotal role as reliable sources of scientific information.Additionally, the analysis delves further into the field, revealing 21 distinct disciplines within the dataset, with the top 10 most prevalent fields illustrated in Figure 13.It is crucial to acknowledge that papers falling into multiple categories, such as higher education and academia, were individually counted as separate occurrences.Additionally, the analysis delves further into the field, revealing 21 distinct disciplines within the dataset, with the top 10 most prevalent fields illustrated in Figure 13.It is crucial to acknowledge that papers falling into multiple categories, such as higher education and academia, were individually counted as separate occurrences.
As depicted in Figure 13, the majority of the papers encompass a broad scope of education and fall under the "general education" category, comprising 104 occurrences.Additionally, our analysis reveals the integration of GAI in various educational frameworks across different fields.Particularly noteworthy is the significant number of papers originating from the medical field, with 43 occurrences, indicating the growing adoption of GAI in both medical and educational contexts.Academia-related papers account for 21 occurrences, further highlighting the relevance of GAI in academic research.Moreover, we observed GAI's intersection with specific disciplines, such as computer science education, nursing education, and engineering education, with thirteen, nine, and nine occurrences, respectively.Lastly, it is worth mentioning the representation of GAI in other fields, including higher education and communication education, with eight and six occurrences, respectively, showcasing the diverse applications and potential of this technology in various educational domains.Additionally, the analysis delves further into the field, revealing 21 distinct disciplines within the dataset, with the top 10 most prevalent fields illustrated in Figure 13.It is crucial to acknowledge that papers falling into multiple categories, such as higher education and academia, were individually counted as separate occurrences.Moving forward, we conducted an analysis of the AI tools utilized in each paper, providing valuable insights into the tool landscape within the field of GAI in education (Figure 14).It is important to note that some papers utilize multiple AI tools, with each tool recorded as a distinct occurrence.Particularly noteworthy is the dominance of ChatGPT, which features in a significant number of papers, with a staggering 151 occurrences.This remarkable prevalence highlights the widespread recognition and adoption of ChatGPT within educational contexts, underlining its efficacy in generating coherent and contextually relevant responses for educational applications.As depicted in Figure 13, the majority of the papers encompass a broad scope of education and fall under the "general education" category, comprising 104 occurrences.Additionally, our analysis reveals the integration of GAI in various educational frameworks across different fields.Particularly noteworthy is the significant number of papers originating from the medical field, with 43 occurrences, indicating the growing adoption of GAI in both medical and educational contexts.Academia-related papers account for 21 occurrences, further highlighting the relevance of GAI in academic research.Moreover, we observed GAI's intersection with specific disciplines, such as computer science education, nursing education, and engineering education, with thirteen, nine, and nine occurrences, respectively.Lastly, it is worth mentioning the representation of GAI in other fields, including higher education and communication education, with eight and six occurrences, respectively, showcasing the diverse applications and potential of this technology in various educational domains.
Moving forward, we conducted an analysis of the AI tools utilized in each paper, providing valuable insights into the tool landscape within the field of GAI in education (Figure 14).It is important to note that some papers utilize multiple AI tools, with each tool recorded as a distinct occurrence.Particularly noteworthy is the dominance of ChatGPT, which features in a significant number of papers, with a staggering 151 occurrences.This remarkable prevalence highlights the widespread recognition and adoption of ChatGPT within educational contexts, underlining its efficacy in generating coherent and contextually relevant responses for educational applications.AI-powered Chatbots, such as BlenderBot and LaMDA, and AI text generators (AITG) appear in the dataset, albeit with lower frequencies [73,110].Tools such as Codex and GitHub Copilot, which translate natural language to code, appear in a limited number of papers [5,154].BERT, a method of pre-training language representations, and Conditional GAN, a type of GAN that involves the conditional generation of images by a generator model, appear a few times in the dataset.Other unique tools with a low number of occurrences include EDU-AI, a model for generating adaptable and functional classroom layouts [94], MACHE-bot, an AI dialogue system [96], and Djehuty (AI), an AI-powered educational gamified environment [97].Furthermore, several papers do not specify a particular tool and instead use more general terms, like "GAI" and "AI", to describe the approaches they employed.AI-powered Chatbots, such as BlenderBot and LaMDA, and AI text generators (AITG) appear in the dataset, albeit with lower frequencies [73,110].Tools such as Codex and GitHub Copilot, which translate natural language to code, appear in a limited number of papers [5,154].BERT, a method of pre-training language representations, and Conditional GAN, a type of GAN that involves the conditional generation of images by a generator model, appear a few times in the dataset.Other unique tools with a low number of occurrences include EDU-AI, a model for generating adaptable and functional classroom layouts [94], MACHE-bot, an AI dialogue system [96], and Djehuty (AI), an AI-powered educational gamified environment [97].Furthermore, several papers do not specify a particular tool and instead use more general terms, like "GAI" and "AI", to describe the approaches they employed.
Then, we outlined the types of research in the context of GAI in education to gain valuable insights into the nature and focus of the scholarly work conducted in this field (Figure 15).Then, we outlined the types of research in the context of GAI in education to gain valuable insights into the nature and focus of the scholarly work conducted in this field (Figure 15).The most prevalent type of research is "General", with 70 occurrences.This suggests a broad range of papers that cover a wide array of topics, methodologies, and perspectives within the domain of GAI in education.Following general papers, "Analysis" emerges as the second most common type, appearing in 47 papers.This indicates a significant emphasis on examining and interpreting data, models, or educational contexts through analytical lenses, allowing for a deeper understanding of the implications and applications of GAI."Assistance" is the third most frequent type, featured in 37 papers.This underscores the relevance of research aimed at developing AI systems that provide support and guidance to learners or educators.It suggests a strong interest in exploring how GAI can serve as an intelligent assistant in educational settings."Learning", "Teaching", and "Testing" papers follow with 21, 18, and 17 occurrences, respectively.These types highlight the focus on investigating how GAI can enhance and optimize various aspects of the learning and assessment processes within educational environments.Moreover, "GAI Threats", "Tutoring", "Research", and "Interaction" types make appearances in a more limited number of papers.These occurrences highlight specific areas of interest, such as addressing potential threats associated with GAI in education, exploring the use of AI systems in tutoring scenarios, conducting research studies, and examining interaction dynamics.

Conclusions and Future Research
AI has emerged as a revolutionary force in the era of rapid technological advancements, profoundly transforming various aspects of human life.In the field of education,

General
Analysis Assistance Learning Teaching Testing GAI Threats Tutoring Research Interaction Type of research The most prevalent type of research is "General", with 70 occurrences.This suggests a broad range of papers that cover a wide array of topics, methodologies, and perspectives within the domain of GAI in education.Following general papers, "Analysis" emerges as the second most common type, appearing in 47 papers.This indicates a significant emphasis on examining and interpreting data, models, or educational contexts through analytical lenses, allowing for a deeper understanding of the implications and applications of GAI."Assistance" is the third most frequent type, featured in 37 papers.This underscores the relevance of research aimed at developing AI systems that provide support and guidance to learners or educators.It suggests a strong interest in exploring how GAI can serve as an intelligent assistant in educational settings."Learning", "Teaching", and "Testing" papers follow with 21, 18, and 17 occurrences, respectively.These types highlight the focus on investigating how GAI can enhance and optimize various aspects of the learning and assessment processes within educational environments.Moreover, "GAI Threats", "Tutoring", "Research", and "Interaction" types make appearances in a more limited number of papers.These occurrences highlight specific areas of interest, such as addressing potential threats associated with GAI in education, exploring the use of AI systems in tutoring scenarios, conducting research studies, and examining interaction dynamics.

Conclusions and Future Research
AI has emerged as a revolutionary force in the era of rapid technological advancements, profoundly transforming various aspects of human life.In the field of education, the emergence of GAI has sparked a transformative revolution, influencing areas such as medical and engineering education.GAI's versatile applications include assessment, personalized learning support, and intelligent tutoring systems.The integration of GAI in education has prompted discussions on ethics, academic integrity, and its potential to reshape teaching and learning methods.As GAI continues to advance, thorough research on its applications, implications, and challenges becomes essential in shaping the future of education.This review paper comprehensively analyzed the research on GAI in education, synthesized key findings and insights from 207 publications, and identified research gaps and future directions in the field.
The first step of this work was to conduct a comprehensive examination of the utilization of GAI in education through an in-depth study and content analysis of the literature in the field.
The content analysis has provided significant findings that highlight the transformative impact of GAI in education.Across various domains and contexts, the integration of GAI, including AI chatbots and virtual instructors, offers substantial opportunities for enhancing educational practices and improving learning outcomes.In computer science education, GAI has proven to be a powerful tool in teaching and learning programming concepts.The research papers revealed the potential benefits of AI technologies for creative programming education and the use of AI-powered code generation tools, like Copilot.Engineering education also benefits greatly from GAI's integration, empowering students to generate innovative solutions through advanced chatbots and text-generation models.Cloud-based frameworks and social robots enhance interactive learning experiences, creating a technologically advanced and effective engineering education landscape.Medical education sees promising applications of GAI in fields like medicine, dentistry, pharmacy, and public health education.Nursing education benefits from ChatGPT in enhancing patient care and fostering critical thinking skills among nursing students.In communication education, GAI models showcase potential in journalism, media education, and healthcare communication.The studies demonstrate GAI's effectiveness in enhancing translation accuracy, emphasizing the importance of educating students about responsible AI use and ethical considerations.In academia, GAI offers potential benefits in various academic fields, such as using AI technologies, natural language processing techniques, and LLMs for research.Furthermore, the subsection on "General Education" provided a comprehensive overview of GAI's wider applications in the educational landscape.Diverse research papers explored technology adoption, classroom design, assessment practices, language instruction, ethics, subject-specific education, learning outcomes, student engagement, challenges in AI integration, and innovative pedagogical approaches.
After conducting the content analysis, the next step involved a bibliometric analysis of GAI within the context of education.It included an examination of a total of 437 pre-existing articles and publications retrieved from the Scopus database.After conducting a literature screening and removing duplicates, the analysis comprised a list of 217 publications that focused on GAI in education.
The comprehensive bibliometric analysis of GAI in education yielded several key findings.The research on GAI experienced significant and exponential growth from 2018 to 2023, with a notable surge in papers published in 2023, which was likely due to the popularity and innovation sparked by tools like ChatGPT.In addition, ChatGPT emerged as the dominant AI tool in GAI research within educational contexts, indicating its widespread adoption for intelligent interactions and contextually relevant responses.The interdisciplinary nature of GAI research in education, spanning computer science, engineering education, and medical education, highlights the need for collaboration in addressing complex challenges and potential applications.
While highlighting the promising contributions of GAI integration in education, this review also delved into potential concerns associated with its implementation.This paper navigated the landscape of ethical implications, responsible GAI use, data privacy safeguards, biases, and academic integrity when exploring the assimilation of GAI within educational settings.These multifaceted concerns were considered in various sections of this review paper.
Based on the findings from this review paper, several future research directions can be suggested to advance the field of GAI in educational settings:

•
Exploring Transparency: Future research should focus on enhancing the transparency of GAI models.Understanding how AI-generated outputs are generated and providing transparent explanations to users can increase user trust and acceptance of GAI tools in educational settings;

Figure 2 .
Figure 2. The number of papers and publications in relation to GAI in education (2018-2023).

Figure 3 .
Figure 3. Summary of GAI research and applications in education.

Figure 3 .
Figure 3. Summary of GAI research and applications in education.

Figure 4 .
Figure 4. Co-occurrence map of text data.

Figure 4 .
Figure 4. Co-occurrence map of text data.

Figure 6 .
Figure 6.Co-occurrence map of all keywords.

Figure 6 .
Figure 6.Co-occurrence map of all keywords.

Figure 7 .
Figure 7. Co-occurrence map of author keywords.

Figure 8 .
Figure 8. Co-occurrence map of index keywords.

Figure 11 .
Figure 11.Bar graph of the top 10 sources by number of publications.

Figure 11 .
Figure 11.Bar graph of the top 10 sources by number of publications.

Figure 13 .
Figure 13.Occurrences of unique fields in education.

Figure 15 .
Figure 15.Occurrences of research types in education.

Figure 15 .
Figure 15.Occurrences of research types in education.

Table 2 .
The application of GAI in engineering education.

Table 7 .
The application of GAI in academia.

Table 8 .
Application of GAI in general education.
• Review of artificial intelligence in education.•ChatGPT'srolesandchallenges in early childhood education.•Opportunitiesandchallenges of large language models, including ChatGPT, in education.•Pi-Mindagent,acognitive digital clone designed to emulate human decision-making processes.•Teacher-studentnetwork framework that combines GAN and Variational Autoencoder (VAE) to enable lifelong learning.

Table 9 .
Top 10 terms by occurrences.

Table 9 .
Top 10 terms by occurrences.

Table 11 .
Top 10 countries by link strength.

Table 12 .
Top 10 authors by documents produced.

Table 13 .
Top 10 authors by citations.
Figure 10.Co-occurrence map of authors.

Table 12 .
Top 10 authors by documents produced.

Table 13 .
Top 10 authors by citations.
• Addressing Biases and Fairness: As GAI models learn from existing data, they may inherit biases present in the data, potentially leading to biased outputs.Future research should focus on addressing and mitigating biases in GAI tools, especially in educational contexts, to not perpetuate stereotypes or discriminate against certain groups of learners; • Designing AI-Enhanced Curriculum: Researchers can explore the potential of GAI in generating adaptive and dynamic learning materials that cater to individual learners' needs.AI-powered curriculum design can foster personalized and engaging learning experiences, promoting lifelong learning and skill development; • GAI in Teacher Professional Development: Research can investigate how GAI tools can assist educators in improving their teaching practices, developing tailored instructional materials, and providing real-time feedback on their performance; • Collaborative AI in Education: Research the potential of collaborative AI systems in education, where humans and AI work together to achieve educational goals; • Longitudinal Studies: Conduct longitudinal studies to track the long-term effects of GAI integration in education.These studies can provide insights into the sustained impact of GAI on learning outcomes, retention rates, and academic performance over extended periods; • Privacy and Data Security: Research should focus on developing robust data protection measures and ensuring that student data is handled responsibly and securely; • Long-term Impact on Learning Outcomes: Investigate the long-term impact of GAI integration on learning outcomes, academic achievement, and students' problemsolving skills; • Ethical Considerations and Responsible AI: Further explore the ethical implications of using GAI in education, addressing concerns related to plagiarism, academic integrity, and the potential impact on students' critical thinking skills.Develop guidelines and policies to ensure the responsible and ethical use of GAI technologies in educational settings; • Student Acceptance and Adoption: Conduct research on students' acceptance and adoption of GAI technology in the learning process.Understand factors influencing their attitudes toward AI-powered tools and identify strategies to enhance student engagement and acceptance; • Interdisciplinary Collaborations: Foster interdisciplinary collaborations between educators, AI researchers, and policymakers to develop comprehensive frameworks for integrating GAI into education; • Inclusivity and Accessibility: Explore ways to make GAI-powered educational tools more accessible to diverse learners, including those with disabilities or language barriers; • AI in Assessment and Evaluation: Investigate the effectiveness of GAI in assessment practices, including essay grading, problem-solving evaluation, and personalized feedback; • AI and Pedagogy: Investigate the implications of GAI for pedagogical practices, examining how it can complement and enhance traditional teaching methods.