International Journal of Educational Research Open 8 (2025) 100401

Contents lists available at ScienceDirect

International Journal of Educational Research Open

journal homepage: www.elsevier.com/locate/ijedro

Challenges of implementing ChatGPT on education: Systematic

literature review
Iván Miguel García-López a , Carina Soledad González González b,
María-Soledad Ramírez-Montoya c,d , José-Martín Molina-Espinosa a,d,*
a
School of Engineering and Sciences, Tecnologico de Monterrey, Monterrey N.L., Mexico
b
University of La Laguna, Tenerife, Spain
c
EGADE Business School, Tecnologico de Monterrey, Monterrey N.L., Mexico
d
Institute for the Future of Education, Tecnologico de Monterrey, Monterrey N.L., Mexico

A R T I C L E I N F O A B S T R A C T

Keywords: Since its launch in 2022, ChatGPT has sparked considerable interest in higher education, raising debates about its
Higher education benefits, challenges, and ethical implications. This systematic literature review, spanning January 2019 to
Educational innovation January 2024, analyzes 42 articles from Web of Science and Scopus to identify key opportunities and challenges
Chatgpt
in its academic integration. Four core issues emerge: (a) technological integration and obsolescence, emphasizing
Generative artificial intelligence
Education 4.0
the need for scalable, modular infrastructures; (b) personalization and equity, focusing on the balance between
individualized learning and avoiding algorithmic bias; (c) data quality and security, highlighting the importance
of transparent data management and robust encryption to protect sensitive information; and (d) ethics and
human-AI collaboration, stressing the importance of institutional policies and continuous teacher intervention to
ensure responsible and effective use. This study advances the discourse by recommending sustainable strategies
for AI adoption, including professional development and fairness audits, while underscoring the critical role of
human oversight in maximizing ChatGPT’s educational impact. Ultimately, it offers actionable insights for in­
stitutions to align AI use with ethical principles and long-term educational goals.

1. Introduction birth of new challenges and opportunities in the area.

Trademarks appear to put a face to these technological resources.
Artificial intelligence (AI) can be applied to almost all human ac­ Currently, the most popular generative artificial intelligence chatbot is
tivities, which has increased its acceptability (Nadarzynski et al., 2020). ChatGPT. It can generate unpublished content and immediate responses
AI-based applications provide access to datasets, which has proven due to its information processing (Mattas, 2023). It has >20 million
helpful for research, spurring new opportunities and challenges for re­ conditions, making it the most potent chatbot over its competitors, such
searchers, companies, and service providers (Firat, 2023). The latter, as Google’s Gemini (Liesenfeld, 2023). Its most outstanding functions
such as Amazon or OpenAI, invest considerable sums in AI to generate include (a) maintaining a conversation with the user (natural language)
solutions to support productivity. AI offers computing capabilities that and (b) responding to almost any topic as long as the information is in its
generate human-like responses (Goralski & Tan, 2022). Traditionally, AI database (Koubaa, 2023). There are still limitations to using this
only solved problems but did not generate new information (Dwivedi resource type, as they are still in the launch phase. However, they will
et al., 2023). After the health contingencies of 2020, AI evolved due to make it possible to change the course of different areas if the contrary
(a) interest in this field of study (Lim et al., 2022), (b) the evolution of opinions about ChatGPT do not hinder it.
new computing capabilities (Buhalis et al., 2023), and c) the imple­ There are mixed opinions on the challenges and benefits of incor­
mentation of neural networks (Lundmark, 2022). With the development porating artificial intelligence into higher education. Consistent with
of AI, technological resources such as Chatbots allow students to ask previous research, chatbots such as ChatGPT can potentially transform
questions and get answers (Buhalis et al., 2023). The above entails the education in several ways (AlAfnan et al., 2023; Kasneci et al., 2023).

* Corresponding author.
E-mail addresses: [email protected] (I.M. García-López), [email protected] (C.S. González González), [email protected] (M.-S. Ramírez-Montoya), jose.
[email protected] (J.-M. Molina-Espinosa).

https://doi.org/10.1016/j.ijedro.2024.100401
Received 4 May 2024; Received in revised form 21 October 2024; Accepted 4 November 2024
Available online 20 November 2024
2666-3740/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

However, significant challenges need to be addressed (Villarreal, 2023). ethically questionable (Dignum et al., 2021). To take advantage of
This paper is conceptual. This article aims to identify the literature on ChatGPT’s personalization capabilities without compromising privacy,
artificial intelligence in education, highlighting ChatGPT as a generative institutions can use anonymized or pseudonymized data (Huang, 2023).
learning resource and the challenges involved in implementing this Additionally, it is important to design systems that allow students to
technology in academic institutions; this is the main differentiator from control what data is collected and how it is used (Zhu et al., 2023).
similar articles. The rationale is to identify some gaps in the education Transparency policies in data collection and use are equally crucial to
literature and provide suggestions for future research, particularly on maintaining the trust of students and parents. However, the rapid evo­
the challenges faced by higher education institutions (HEI) regarding lution of technology poses another significant challenge: managing
automating activities, clarifying concepts, and generating personalised technological obsolescence.
content. The rapid evolution of technology presents significant practical
challenges, such as the need to continuously update education systems
1.1. Rationale for the systematic review to keep up with the latest innovations. This can be costly and logistically
complex (Sallam, 2023), especially for institutions with limited re­
Although there are several reviews on Generative Artificial Intelli­ sources (Govea et al., 2023). To mitigate the challenges of technological
gence in education, this work focuses specifically on ChatGPT and its obsolescence, the adoption of modular and scalable technology archi­
unique challenges in higher education, encompassing ethical, techno­ tectures can be an effective solution (Sung et al., 2022). These archi­
logical, and security aspects that have not been thoroughly explored in tectures allow for incremental upgrades without the need for complete
previous studies (Dempere et al., 2023). In examining these challenges, system overhauls. Finally, continuous training of staff in new technol­
key areas that need to be carefully considered when implementing ogies is essential to ensure effective and seamless integration, thus
ChatGPT in educational settings are highlighted (Lo et al., 2024), such as closing the cycle of challenges and solutions that accompany the
data privacy, equity in outcomes, academic integrity, personalization of implementation of ChatGPT in higher education.
learning, and management of technological obsolescence.
One of the most important ethical challenges in the adoption of 1.2. Bridging gaps: how this research advances beyond existing literature
ChatGPT in education is data privacy and security. Since these tech­
nologies handle large volumes of personal data, there is a significant risk This paper focuses on the specific challenges of implementing
that sensitive student information could be compromised (Sharma et al., ChatGPT in higher education institutions, highlighting ethical, techno­
2019; Xu & Ouyang, 2022). This risk is even greater when it comes to logical, and pedagogical challenges. From a comprehensive literature
health data or sensitive information used to personalize the learning review, relevant studies exploring the integration of generative artificial
experience. To mitigate these risks, it is essential to implement strict intelligence tools in education were identified. However, this research
data protection policies and ensure that all interactions with AI comply differs by delving into aspects such as technological obsolescence and
with international regulations (Zhang & Tur, 2024) such as the GDPR in the need for institutional policies that ensure sustainable and adaptive
Europe (Michel-Villarreal et al., 2023). In addition, advanced encryp­ integration (Noroozi et al., 2024). This distinction lays the groundwork
tion and authentication techniques must be employed to protect data for analyzing the gaps found in prior research, setting the stage for a
both in transit and at rest (Rahman, 2022). In turn, it is crucial that more comprehensive approach to AI implementation.
educational staff receive training in cybersecurity to minimize possible Among the works consulted, the study by Noroozi et al. (2024) ad­
security breaches. However, privacy is not the only challenge; The issue dresses the pedagogical conditions necessary to integrate ChatGPT,
of algorithmic bias must also be addressed. highlighting student motivation and personalization of learning while
AI models, such as ChatGPT, can perpetuate existing biases in also identifying ethical risks, such as algorithmic biases. This research
training data, which could result in discriminatory outcomes, especially complements those findings by emphasizing the importance of a robust
in educational contexts. This bias has the potential to negatively affect infrastructure to overcome the risks of technological obsolescence,
groups of students who are already disadvantaged, exacerbating current ensuring continuous and sustainable integration. Whereas Al-Maskari
inequalities (Holstein & Doroudi, 2021). To address this problem, it is et al. (2024) focus on the immediate benefits of learning, this work
critical to train these models with diverse and representative datasets places greater emphasis on how to personalize education without
(Gödde et al., 2023) that reflect the variety of students’ experiences and compromising long-term equity (Ali et al., 2024). This shift from
contexts (Guidotti et al., 2018). It is also important to implement equity short-term educational outcomes to long-term sustainability is key for
audits to identify and correct biases in the results generated by AI advancing institutional practices.
(Bender et al., 2021). Transparency in algorithms and access to clear Another relevant aspect identified in the literature is the importance
explanations of decisions made can also help mitigate these risks. With of human intervention in the use of AI tools, as suggested by Lo et al.
equity in mind, it is crucial that education policies also protect academic (2024). Their work highlights that teacher supervision is essential to
integrity. avoid errors and improve pedagogical effectiveness. However, this study
The use of ChatGPT in education raises concerns about academic extends that perspective by pointing out the need for continuous
integrity, as students could use AI to generate answers or completed training for educators, ensuring their adaptation to these emerging
essays, putting the originality and authenticity of their academic work at technologies and promoting long-term ethical and effective use (Aler
risk. To safeguard academic integrity, institutions must develop clear Tubella et al., 2024). This connection between technology and profes­
policies that define the appropriate use of AI in education (Holmes et al., sional development emphasizes the need for dynamic strategies that
2023; Xu & Ouyang, 2022). In addition, it is important to educate stu­ enable teachers to thrive in evolving educational landscapes.
dents about the importance of academic honesty and the ethical impli­ In terms of ethical risks, concerns have been identified around
cations of using AI inappropriately (Liebrenz Rosenstock et al., 2023). algorithmic biases and privacy issues, which could affect trust in these
Advanced plagiarism detection tools can also be useful in identifying platforms if not properly managed (Tlili, 2023). To address these chal­
unauthorized use of AI in academic papers (Lo, 2023). Beyond academic lenges, the implementation of fairness audits and the use of transparent
integrity, another crucial aspect is ChatGPT’s ability to personalize algorithms are proposed, ensuring fair and accessible personalization in
learning. educational settings. This combination of measures seeks to balance
ChatGPT’s ability to personalize learning to individual student needs opportunities and risks, strengthening trust in institutions adopting AI.
is one of its biggest advantages. However, this personalization requires As fairness becomes a cornerstone for acceptance, these strategies
the collection and analysis of large volumes of personal data ensure that technological adoption aligns with ethical principles.
(Vargas-Murillo et al., 2023), which can be logistically challenging and The review also reveals that a modular infrastructure is essential to

2
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

meet the challenges of AI adoption in education. Kurban and Şahin 2.2. Uses of ChatGPT
(2024) emphasize that this management depends not only on techno­
logical infrastructure but also on the creation of sound regulatory Technological resources have quickly adopted the use of GAI for
frameworks to govern its use. Additionally, Tlili (2023) highlight that their operation. Dwivedi et al. (2023) mention that despite the existence
resistance to change is a common barrier to AI adoption, underscoring of other programs that implement GAI, OpenAI’s ChatGPT is the most
the importance of institutional strategies that foster the acceptance and discussed regarding technological innovations due to its machine
effective use of these tools. This interconnectedness between regulation, learning (Damar 2022), i.e., it is continuously updated with new infor­
infrastructure, and institutional culture ensures that AI adoption is not mation (Chatterjee & Dethlefs, 2023). The advent of this chatbot has
only feasible but also meaningful. brought challenges due to its ease of use and affordability, unlike other
This work also brings a fresh perspective by focusing on automating costly chatbots (Yeo-Teh & Tang, 2023). ChatGPT has come to the
academic processes and generating personalized content, integrating AI attention of AI researchers (Michel-Villarreal et al., 2023) to understand
without compromising academic integrity. Unlike studies that focus on its use and limitations, so it does not currently pose a danger to students
formative assessment and lesson plan creation, this research delves into (Su & Yang, 2022). Recent studies suggest that human intervention
how to manage technological obsolescence and ensure equity in higher maximizes the impact of ChatGPT in education (Lo et al., 2024). For
education (Zhang & Tur, 2024). This shift from a narrow focus on example, teachers can monitor AI-generated responses to improve con­
content generation to a broader emphasis on sustainability reflects the tent accuracy and better guide student learning. There is still a long way
growing need for institutions to future-proof their practices. to go in academia, considering that implementing this type of technol­
Overall, this research provides a broad perspective on ChatGPT ogy is just beginning broadly.
implementation in higher education, highlighting that AI adoption must The use of this technological resource is not limited to support for
be managed through clear policies, scalable infrastructures, and ethical research. Its primary proven uses include (a) design of research ques­
frameworks that ensure equity and transparency (Adel et al., 2024). In tions, (b) mathematical problem-solving, and (c) writing of various
addition, human intervention, along with continuous teacher training, is documents (Kasneci et al., 2023). Because it is an open-source GAI tool,
critical to ensure effective and sustainable integration. The imple­ it can perform various tasks (Ilieva et al., 2023), such as answering
mentation of ChatGPT, far from being an exclusively technological questions in many areas. Therefore, it contributes to generating
process, requires a holistic approach that balances educational benefits knowledge (Floridi & Chiriatti, 2020). ChatGPT is trained with a rein­
with risk management and the promotion of equity. This conclusion ties forcement learner where inputs produce subsequent outputs, meaning it
together the various strands of the discussion, emphasizing that sus­ is a pre-processed generative transformer (predictor of responses)
tainable AI integration relies on a blend of technical, pedagogical, and (Dowling & Lucey, 2023). ChatGPT disruptively affects users of educa­
ethical considerations. tional technology, which has generated uncertainty (Zhai, 2022). One of
the benefits of ChatGPT is that it facilitates access to information; for
2. Theoretical framework example, students can write articles and summarise literature. This en­
tails an evolution of current educational practices where information is
2.1. AI support technologies increasingly available to users, not only in a query but also in a syn­
thesised form, ready for use.
Thanks to assistive technologies such as deep learning or natural
word processing, AI has evolved to understand the user better. A sig­ 2.3. Related ChatGPT work
nificant development supporting the growth of AI has been the LLM
(large language models), which are trained on volumes of data from The education sector widely discusses ChatGPT due to its trans­
diverse areas (Ren et al., 2023). Particularly noteworthy is the launch of formative ability to change the methods of conducting and scoring as­
ChatGPT, which is a combination of two words: chat because it is a sessments. Unlike other educational technologies that can be tracked
chatbot and GPT (generative pre-trained transformer), which uses when used, the capabilities of ChatGPT make its use virtually untrace­
learning-through-reinforcement of human feedback with user-guided able (Deng & Yu, 2023). Faced with the dilemma of the teacher’s
refinement (Stiennon et al., 2020), (Chen et al., 2021). It can also fill possible replacement by this type of technological resource, educators
in extremely masked text (Sun et al., 2019) or generate stories from confront the need to implement policies to improve the quality of life
user-defined elements and styles (Pau et al., 2019). It has the ability to using this type of GAI (Bird et al., 2019). This technological resource
be trained by users to improve their performance, which is the primary continues to evolve with increasingly sophisticated functions (Shaw,
"secret" of ensuring its correct functioning and evolution. 2020). However, a nascent disadvantage of using chatGPT is that stu­
Capabilities that are currently recognised and valued include the dents minimise their ability to develop cognitive skills in their tasks.
emulation of a conversation. Conversations in ChatGPT are produced Likewise, the resource can reduce the quality of the work, complicating
thanks to the GPT language model (Terwiesch, n.d.) that interprets user its evaluation.
requests and satisfies needs by processing information to generate new Little by little, more researchers are venturing out to test this tool in
data depending on the information stored in its database. In addition, it their daily activities. Examples include DiGiorgio and Ehrenfeld (2023),
relies on using natural language (LN), that is, the ability to maintain a who used ChatGPT to analyse medical records. (Iftikhar, 2023) analysed
fluid conversation and simulate human conversation with a user, the chatbot as a potential virtual doctor. (Mijwil et al., 2023) touched on
although sometimes with notable errors like word confusion. However, the topic of cybersecurity in Data Protection in the medical sector. (Chiu
the most relevant error lies in writing, for example, being unable (in et al., 2023) stated that some companies use the resource to modernise
October 2023) to differentiate the speaker’s gender (Pavlik, 2023). A and automate their operations like shipping, electronic mail, or report­
generative model creates new data rather than relying on input data ing (George & George, 2023). Okey et al. (2023) looked at ChatGPT’s
(Michel-Villarreal et al., 2023). It is fascinating how the machine learns impact on participants’ emotions, analysing their opinions on social
from data management to improve its responsiveness because it is not networks. Biswas (2023) addressed the possible and future use of
only a matter of continuous use to discover its limitations and benefits ChatGPT in the military sector, such as automated target reconnais­
but also converting today’s errors into complete functionality in the near sance, unmanned vehicles, and missile guidance. Charan et al. (2023)
future. compared the performance of ChatGPT against the Google chatbot for
answering human questions by analysing the complexity of the answers.
It turned out then that its applicability is broad, and only time will tell
what types of tasks it best fits.

3
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

2.4. Ethics using GenAI harmoniously can be fostered (Dwivedi et al., 2023). Thus, the cycle of
an ethical and safe approach in the implementation of ChatGPT in ed­
The use of ChatGPT in education involves handling large volumes of ucation is closed, ensuring both the protection of students’ rights and
personal data, which poses significant risks in terms of privacy and se­ equity in access to new technologies.
curity. Therefore, it is crucial to conduct a thorough Privacy Impact
Assessment (PIA) before implementing this technology (Nguyen et al., 3. Method
2022). This process not only makes it possible to identify the areas
where personal data is most exposed (Holmes et al., 2023) Not only do The methodological approach used in this study is the Systematic
they have to do so, but also develop specific strategies to protect them Literature Review (SLR). This process included the generation of ques­
appropriately (Watanabe, 2023). Once these areas of risk have been tions, the exploration of bibliographic sources, the definition of inclu­
identified, it is essential to implement effective measures to safeguard sion and exclusion criteria, the selection process and data analysis
data privacy such as minimizing the amount of data collected. (Kitchenham et al., 2007). The purpose of the SLR is to identify, analyse,
Among these measures, data minimization involves collecting and and interpret the results of research available in a specific period related
processing only the minimum amount of information necessary to ach­ to the topic of interest (García-Peñalvo, 2021), which in this case
ieve educational goals, thus reducing the risk of personal data exposure. focused on the application of ChatGPT in the educational field. Fig. 1
Anonymization, which removes or modifies personal identifiers, is illustrates the steps followed in this study.
another key strategy that protects students’ identities and mitigates the As a first step, it was necessary to formulate the research questions to
impact of potential data breaches (Crawford, 2022). It is critical for give direction to the research. The questions are presented below in the
institutions to implement strict data retention policies (Floridi & Chir­ subsection corresponding to phase 1.
iatti, 2020), ensuring that personal information is deleted or anony­ Phase 1: Research Questions
mized when it is no longer needed (Bender et al., 2021). However, data Derived from the process selected to develop this SLR, the authors
privacy is not the only challenge; Algorithmic bias must also be began by considering those questions that would obtain the most rele­
addressed and equity in education must be ensured. vant data to enrich the research, mainly through exploring article con­
Algorithmic bias can be mitigated through the use of diverse and tent data besides the statistical data. Table 1 presents four research
representative datasets that reflect a wide range of cultural experiences questions to identify publications on using GAI in higher education
and contexts. It’s essential that models like ChatGPT are trained on data during 2019–2024. These questions and corresponding themes arose
that includes perspectives from different demographics (Buolamwini & due to the review of previous work in this field (Kitchenham, 2007). The
Gebru, 2018), including racial, gender, and socioeconomically diverse development of the questions came from identifying opportunities in the
communities (Barocas et al., 2018). This diversity in training data helps existing literature and the experiences and challenges present in the
reduce the possibility of perpetuating existing biases in society (Bender academic community. The possible answers were derived from the
et al., 2021; Mitchell et al., 2019). However, even with these efforts, it is theoretical basis of the research. The motivation to bring new insights to
crucial that critical decisions in education are not fully automated. the study of ChatGPT implementation in higher education drove the
To mitigate the impact of algorithmic biases, it is critical that creation of the research topics and questions.
important decisions have human intervention. The inclusion of human These questions defined the elements of the articles that shaped the
feedback in the decision-making process allows errors to be identified research work, especially trying to discover crossovers in elements such
and corrected (Floridi & Chiriatti, 2020) or biases before they affect as the challenges and barriers of using ChatGPT in education.
students (Raji et al., 2020). Building cross-functional teams, comprised Phase 2: Search Process
of educators, ethicists, and AI developers, is key to ensuring that The exploration of documents occurred through querying Scopus and
AI-based decisions are fair and equitable (Crawford, 2022; Liebrenz Web of Science (WoS) due to their comprehensive coverage. The criteria
Rosenstock et al., 2023). As we move forward with AI integration, we used were keywords (ChatGPT and challenges in education), time in­
must also protect academic integrity. terval (2019–2024), and document type (article). The information came
To safeguard academic integrity, institutions must establish clear fromtheir databases on January 15, 2024. Table 2 shows the search
policies that define the limits of the use of ChatGPT and other AI tech­ sequence used in the research process through WoS and Scopus.
nologies in education. These policies should specify how students can Phase 3: Inclusion and Exclusion Criteria
use these tools ethically (Aler Tubella et al., 2024) and the consequences In conducting research on the application of ChatGPT in educational
of misuse, such as plagiarism or fabrication of content (Raji et al., 2020). contexts, it is crucial to ensure that the selected literature accurately
It is crucial that these guidelines are clearly communicated to all stu­ reflects the focus of the study. ChatGPT has various potential uses in
dents and teachers to ensure a common understanding (Baena-Rojas education, ranging from assisting in solving tasks to explaining complex
et al., 2023; Fan, 2023). In addition to policies, tools that reinforce these concepts and addressing educational challenges. Given the broad scope
principles need to be adopted. of this topic, the authors aimed to refine their search to focus explicitly
Institutions should implement advanced plagiarism detection tools on the intersection of ChatGPT and its challenges within educational
capable of identifying the unauthorized use of AI in academic papers. settings.
These tools should be designed to recognize AI-generated writing pat­ To achieve this, a global search was conducted using the term
terns (Alexander et al., 2023) and compare them with previously sub­ "ChatGPT," with a specific emphasis on "challenges in education" within
mitted works (Jarrah et al., 2023). Integrating these tools into learning a defined period from 2019 to 2024. This timeframe was selected to
management systems can help proactively monitor and prevent plagia­ capture the most recent and relevant developments in the field,
rism (Chiu et al., 2023). Finally, it is critical for students to understand acknowledging that the rapid evolution of AI technologies necessitates
the ethical implications of using AI in their academic work, which brings an up-to-date analysis.
us to the need for digital ethics education. Justification for Keywords and Database Search:
To create a culture of integrity and accountability in the use of The careful selection of keywords such as "ChatGPT" and "challenges
technological tools in education, institutions must offer workshops and in education" is essential for narrowing down the vast body of literature
courses on digital ethics. These programs should address issues such as to those studies most pertinent to the research questions. Keywords play
academic honesty, responsible use of technology (Aler Tubella et al., a critical role in database searches because they directly influence the
2024) and the consequences of academic misconduct (Gulati et al., relevance and accuracy of the results. By ensuring that these terms were
2024). Through this education in digital ethics, an educational envi­ present in the titles, abstracts, or keywords of the articles, the authors
ronment where integrity and technological innovation coexist maximized the likelihood of retrieving studies that specifically address

4
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

Fig. 1. SLR phases.

Source: Own authorship based on Kitchenham (2007).

Table 1 Table 2
Research questions. Search strings.
Topics Research Questions (RQ) Possible Answers Search string in Scopus Search string in WoS

Focuses on the overall RQ1. Under what Pedagogical Integration TITLE-ABS-KEY (chatgpt AND challenges chatGPT (All Fields) and challenges and
conditions and conditions and through (Hasanein & Sobaih, AND education) AND (LIMIT-TO education (All Fields) and Article
strategies for what specific strategies 2023) (SUBJAREA, "SOCI")) AND (LIMIT-TO (Document Types) and 2019
integrating AI tools like can generative AI tools Technological (DOCTYPE, "ar") AND (LIMIT-TO (Publication Years) and 2020
ChatGPT in higher like ChatGPT be Infrastructure (Hwang (PUBYEAR, 2019) OR LIMIT-TO (Publication Years) and 2021
education. effectively integrated into et al., 2020). (PUBYEAR, 2020) OR LIMIT-TO (Publication Years) and 2022
higher education to Faculty Training and (PUBYEAR, 2021) OR LIMIT-TO (Publication Years) and 2023
enhance learning Support (Hasanein & (PUBYEAR, 2022) OR LIMIT-TO (Publication Years) and 2024
outcomes while Sobaih, 2023) (PUBYEAR, 2023) OR LIMIT-TO (Publication Years)
maintaining ethical Ethical Guidelines and (PUBYEAR, 2024)
standards and addressing Policies (Sharma et al.,
operational challenges? 2019) Source: Authors.
Student Engagement and
Accessibility (Xu & aligned with the research objectives.
Ouyang, 2022)
The inclusion criteria mandated that articles must contain both
Challenges in the RQ2 How do • User challenges
Implementation of technological • Operational "ChatGPT" and "challenges in education" in their titles, abstracts, or
ChatGPT in Education: obsolescence and challenges keywords, ensuring a direct connection to the research focus. Addi­
Users, Operations, integration challenges • Environmental tionally, the timeframe of publication between January 2019 and
Environment, affect the ability of Challenges January 2024 was selected to include the most current research,
Technology and Ethics. educational institutions to • Technological
keep their education challenges
reflecting the ongoing developments and challenges in applying
systems up-to-date with • Ethical challenges ChatGPT in education. This period captures the advent and subsequent
the latest technologies, (Ali et al., 2024) rise in the adoption of ChatGPT, providing a comprehensive view of its
and what are the most impact and the issues it raises in educational contexts.
effective strategies to
Exclusion Criteria and Rationale:
mitigate these issues?
The Implementation of RQ3 In what ways can AI • Ethical Issues To maintain the study’s relevance and rigor, certain document types
ChatGPT in Education: technologies like • Generalization and were excluded, including book chapters, books, and reviews. These
Challenges, Ethical ChatGPT personalize adaptation sources often provide broader overviews or synthesize existing research
Considerations, and Key education to fit the needs • Evaluation and rather than offering new empirical insights, which could dilute the focus
Aspects of a diverse student base, feedback
and what ethical and • Policy Development
on specific challenges related to ChatGPT in education. Additionally,
practical challenges arise • Customization and articles published before January 2019 or after January 2024 were
when trying to minimize Scalability excluded to ensure that the research is both contemporary and relevant,
educational inequities (Bandi et al., 2023) avoiding outdated perspectives or speculative future scenarios that may
through personalization?
not reflect the current state of the field.
Limitations and RQ4 What are the main • Limited access to real-
restrictions to ensure challenges associated time data By employing these inclusion and exclusion criteria, the authors
proper use and with data quality and • Dependence on input ensured that the selected literature is not only relevant but also directly
interpretation of security in education data addresses the unique challenges of integrating ChatGPT into educational
ChatGPT-generated systems that use AI, and • Token limitations environments. This methodological rigor allows for a focused explora­
results what measures can be • Data Synchronization
implemented to ensure Challenges
tion of how ChatGPT can both positively and negatively impact educa­
data integrity and • Input Data Reliability tion, providing a solid foundation for further analysis and discussion.
protection? (Alshami et al., 2023) Phase 4: Data selection and extraction process
Source: Authors. We found 51 articles in Scopus and 14 in WoS and removed 8 du­
plicates. To ensure that the articles focused on using ChatGPT in edu­
cation as the main topic, we applied quality criteria to verify that the
the intersection of ChatGPT and educational challenges. This approach
articles on technology and education included “GAI” in their titles, ab­
not only improves the efficiency of the literature review but also en­
stracts, or keywords.
hances the quality of the findings by focusing on content that is most

5
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

We excluded 15 articles that did not address using ChatGPT in ed­ methodology proposed by Kitchenham et al. (2010). To summarise the
ucation. Ultimately, 42 articles remained (40 from Scopus and 2 from responses, the authors created the graphical representations found in
Web of Science) to conduct the SLR. The analysis examined data such as this document. The articles’ terms, keywords, and objectives were also
authors’ names, titles, DOIs, abstracts, and countries of origin. Fig. 2 examined to uncover convergences and areas of potential interest.
shows the scheme of the study following the Prisma model. (Page et al.,
2021). 5. Results
Many articles were suppressed because their orientation was not
strictly academic, i.e., they talked about testing ChatGPT from an en­ Initially, evaluating the articles identified the approaches pursued in
gineering aspect. Likewise, they did not meet the quality criteria; for the research, discovering the works’ meaning and identifying new terms
example, the keywords, despite being present in the content of the ar­ and those "going out". This type of review is desirable with an extensive
ticles, did not comply with the focus of this research. database. Fig. 3 shows the word cloud of the articles’ keyword trends.
Concepts continuously evolve, especially when new tools such as AI-
4. Data synthesis based resources appear. Fig. 3 reveals more likely trends in areas directly
involving artificial intelligence, chatGPT, and education. It highlights
We proceeded to examine the data collected from the database of the reduction of concepts such as software, programming, and knowl­
each article to address the four questions of the Systematic Literature edge because, with the arrival of tools such as ChatGPT, the use of
Review (SLR). The qualitative content analysis aimed to draw objective programming became minimized, new terms were adopted instead of
inferences from the information provided. This process occurred in two software, and the focus beyond generating knowledge became how to
phases: a) a mechanical one to structure the data into the corresponding use the AI-generated content. Undoubtedly, new trends will be debated
thematic categories, and b) an interpretive one to identify which data in the coming years.
responded effectively to the research questions, following the Below are the answers to the research questions posed in the

Fig. 2. PRISM.
Source: Based on Page et al. (2021).

6
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

respects the rights of students and maintains trust in these tools.

Finally, inclusion and equity must be fundamental pillars in the
implementation of advanced educational technologies. It is crucial that
all students can participate in and benefit from the use of AI tools,
regardless of their technical background or skills (Holstein & Doroudi,
2021). This requires the provision of resources and support to ensure
that AI-powered education is accessible to all (Antoninis, 2023),
including those with disabilities, thereby fostering greater equity in
access to education (Hwang et al., 2020; Xu & Ouyang, 2022). By
ensuring that all students, without exception, can take advantage of
these opportunities, a more inclusive and equitable education is
promoted.
Fig. 3. Article content word cloud. RQ2 How do technological obsolescence and integration chal­
Source: Authors. lenges affect the ability of educational institutions to keep their
education systems up-to-date with the latest technologies, and
previous section, intending to generate valuable content for research what are the most effective strategies to mitigate these issues?
and practical questions. The integration of ChatGPT into educational settings offers notable
RQ1. Under what conditions and through what specific strate­ advantages but also presents significant challenges, particularly
gies can generative AI tools like ChatGPT be effectively integrated regarding data protection and student privacy. There are heightened
into higher education to enhance learning outcomes while main­ concerns about the management and safeguarding of students’ personal
taining ethical standards and addressing operational challenges? data, especially given the vast amounts of information processed by AI-
In an increasingly digitized world, education cannot be left behind in based platforms. Ensuring transparency in how this data is used and
the adoption of emerging technologies. The integration of AI tools such adhering to regulatory standards are crucial steps to maintain trust and
as ChatGPT into curricula should be aligned with existing pedagogical uphold the privacy rights of users (Vilalta-Perdomo et al., 2023).
methodologies, so that they complement and do not replace traditional Moreover, this technology holds the promise of providing personalized
teaching practices (Leelavathi & Surendhranatha, 2024). An approach support to students with special needs by adapting learning materials to
based on adaptive pedagogy can be especially beneficial (Mai et al., their unique requirements and offering alternative ways of interaction
2024), allowing AI to strengthen critical thinking and problem-solving not possible through traditional teaching methods (Rakap, 2023). As
skills, which in turn improves academic outcomes (Hasanein & such, ChatGPT has the potential to make high-quality education more
Sobaih, 2023; Modran et al., 2024). This balanced integration ensures accessible, adapting to diverse learning styles and needs (Antonio,
that learning is enhanced without losing sight of the essence of tradi­ 2023). Therefore, while ChatGPT stands out as a transformative tool
tional teaching. capable of customizing and enhancing educational experiences, it re­
For this integration to be effective, it is essential that educational quires careful handling to safeguard students’ privacy and maintain
institutions have a robust technological infrastructure. In the age of confidence in digital education platforms. It’s essential that innovation
digitalization, technological infrastructure becomes an essential pillar in education does not come at the expense of safety and equity.
for academic success. It is necessary to ensure continuous hardware and The introduction of technologies like ChatGPT into educational
software updates to prevent technological obsolescence (Chugh et al., contexts brings considerable technological challenges that need careful
2023) and ensure that AI platforms, such as ChatGPT, work effectively management to guarantee both their safety and effectiveness. Fig. 4
and safely in academic environments (Hwang et al., 2020; Kahraman outlines the technological hurdles encompassing issues such as
et al., 2023). With the right infrastructure, an efficient and secure "compatibility, updates, technological infrastructure, cybersecurity, and
learning environment is created, capable of responding to contemporary obsolescence." These are crucial for ensuring that tools like ChatGPT
technological demands. operate safely and effectively within academic settings (Vilalta-Perdomo
However, technology alone is not enough; Continuous training of et al., 2023). A system that is either poorly integrated or lacking in
teachers is equally indispensable in this constantly evolving educational cybersecurity measures could risk exposing sensitive student informa­
context. Continuous training ensures that educators can use these tools tion or create more issues than it resolves. On the front of user chal­
effectively (Ng et al., 2023), covering both technical (Nguyen et al., lenges, factors like "adaptation, training, user interface, user experience,
2022) and pedagogical training (Göçen & Aydemir, 2020). This helps and barriers to use" underscore the necessity for both educators and
teachers integrate AI into their teaching and adapt their methodologies students to embrace new AI-mediated methods of teaching and learning
to new technologies, preparing them to make the most of these inno­ (Abrahão et al., 2021). Effective and ethical adoption of disruptive
vative tools in their pedagogical practices. The preparation of teaching
staff, therefore, becomes a key element for the success of technological
integration.
In addition, in the integration of advanced technologies in education,
it is crucial to consider the ethical implications that they entail. The
implementation of AI tools should be accompanied by clear ethical
guidelines that address issues such as data privacy, academic integrity
(Holmes et al., 2023), and potential biases in AI models. Recent studies
highlight the importance of specific conditions for effective AI integra­
tion in higher education (Lo et al., 2024). In addition, combining
ChatGPT with human intervention has been shown to enhance educa­
tional outcomes (Zhu et al., 2023). This combination not only mitigates
algorithmic errors, but also enhances the personalization of educational
content protecting the rights of students (Nguyen et al., 2022) and
maintaining trust in these technologies (Sharma et al., 2019; Xu &
Ouyang, 2022). Only in this way can it be ensured that the use of AI
Fig. 4. RQ2 ratio.

7
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

technologies like ChatGPT hinges on comprehensive training and while mitigating risks. Such strategies enhance both learning and
user-friendly interfaces (Connolly, 2023). The ethical dimension of teaching, and also prepare students to responsibly engage with emerging
technology adoption involves critical considerations such as "data pri­ technologies in their future professional and personal lives (Antonio,
vacy, algorithmic bias, transparency, informed consent, and fairness." 2023). Furthermore, educators need not only the technical skills to
Technology challenges include obsolescence and lack of infrastructure. utilize these tools but also the critical ability to assess their effectiveness
At the user level, both educators and students face adaptation and and ethical implications in practice (Levine, 2023). Prioritizing profes­
training barriers that hinder the adoption of ChatGPT in educational sional development in this area is essential for educational institutions
settings, making ongoing training in emerging technologies necessary. committed to the responsible and effective integration of advanced AI
Addressing these issues is essential to prevent the exacerbation of technologies into their curricula.
existing disparities using AI in education. Thus, it is imperative for RQ3 In what ways can AI technologies like ChatGPT personalize
educational institutions to deploy ChatGPT and similar tools within a education to fit the needs of a diverse student base, and what
robust ethical framework designed to manage these potential risks ethical and practical challenges arise when trying to minimize
(Levine, 2023). In summary, as educational institutions increasingly educational inequities through personalization?
adopt sophisticated tools like ChatGPT, it is vital that they implement The discussion surrounding Research Question 3 topics in higher
these technologies within a strong ethical framework that prioritizes education, particularly with respect to ChatGPT and generative AI,
data privacy, mitigates algorithmic bias, and enhances the overall user highlights a complex range of challenges and opportunities. These
experience. The exploration of the relationship between AI technologies include the personalization of learning experiences and the ethical
and higher education reveals several promising opportunities and considerations of data usage. As depicted in Fig. 5, educational
notable challenges. One significant opportunity is the potential for personalization through AI is promising, but it also poses ethical risks if
personalized learning at a previously unimagined scale, which could it is based on unrepresentative data. It is key to ensure that systems
transform higher education by providing education that adapts to stu­ adapt content to individual needs without perpetuating existing biases,
dents’ individual needs. However, this also underscores the pressing promoting equity and avoiding discrimination in diverse educational
need for clear policies, guidelines, and frameworks to responsibly contexts that accommodate diverse student needs and preferences (Zhu
incorporate these technologies into the educational landscape, as well as et al., 2023). A pivotal aspect of this approach is the capacity of AI to
empirical research to better understand users’ experiences and percep­ customize educational materials and experiences to suit individual stu­
tions (Vilalta-Perdomo et al., 2023). Regarding the specific impacts of dent requirements, thereby enhancing accessibility and promoting eq­
ChatGPT in higher education, it is evident that this technology can uity in education. This level of personalization proves particularly
greatly enhance students’ writing and critical thinking skills. ChatGPT advantageous for students with disabilities or those in need of alterna­
helps students organize their thoughts more clearly and communicate tive learning methods, helping to equalize educational opportunities
them more effectively, which is particularly beneficial for those for (Levine, 2023). However, the deployment of technologies like ChatGPT
whom English is not their first language. This support reduces cognitive within educational frameworks also introduces significant ethical di­
strain, enabling students to concentrate more on the substance of their lemmas, particularly concerning data privacy and security
writing rather than on the linguistic complexities (Levine, 2023). (Vilalta-Perdomo et al., 2023). The gathering and analysis of extensive
However, the potential for ChatGPT to autonomously generate complete personal data, essential for tailoring educational experiences, require
responses to exam questions or entire academic essays poses a signifi­ stringent security protocols and must comply with privacy laws to
cant threat to the integrity and originality of student work safeguard student information and uphold their trust.
(Vilalta-Perdomo et al., 2023). From an ethical standpoint, ensuring While the opportunities for enhancing education through generative
fairness and preventing bias in AI applications is crucial. technologies are significant, the challenges they introduce cannot be
Educators and technology designers must ensure that tools like overlooked. It is crucial for educational institutions to establish clear
ChatGPT are accessible to all students and that they deliver unbiased, policies and accountable frameworks that foster innovation and
equitable outcomes. This involves training these systems with diverse personalization while safeguarding students and ensuring ethical AI
datasets to reduce biases related to race, gender, or culture that might usage. This requires ongoing investment in research to better compre­
affect the generated content (Connolly, 2023). It is essential that both hend the effects of these technologies and to formulate best practices for
students and faculty are provided with the necessary resources to criti­ classroom integration (Díaz, 2024). The diversity of users and the range
cally assess and question these technologies, maximizing their benefits of learning environments underscore the need for educational

Fig. 5. RQ3 ratio.

8
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

technologies to be inclusive and adaptable. Such flexibility is essential to

ensure that all students, irrespective of their backgrounds, abilities, or
personal situations, have equal access to the educational benefits offered
by generative AI like ChatGPT (Diaz, 2024). However, successfully
deploying these technologies demands a deep understanding of the
educational contexts in which they operate. Institutions must assess not
only the technical features of tools such as ChatGPT, but also their
interplay with the pedagogical and cultural dynamics of each environ­
ment. This highlights the pressing need for empirical research to
examine how students and teachers perceive and utilize this technology
in their unique settings, ensuring its integration fosters effective and
equitable learning (Zhu et al., 2023). Thus, for the adoption of genera­
tive technologies like ChatGPT to be genuinely beneficial and ethical in
education, educational institutions need to focus not just on the tech­
nical aspects but also on how these technologies are thoughtfully and
sensitively incorporated into various pedagogical and cultural contexts, Fig. 6. RQ4 ratio.
adapting to meet the needs and perceptions of all stakeholders involved.
The development and implementation of ChatGPT in higher educa­ prevalent issue in higher education is the variability in data quality
tion must be rooted in strong ethical principles, emphasizing educa­ across different platforms and systems, which can result in inaccurate or
tional equity and justice. It is crucial that policies are designed to avoid biased outcomes if not properly managed (Dempere et al., 2023).
exacerbating existing inequalities and to foster genuine educational in­ Additionally, data preprocessing and the selection of appropriate data
clusion. By doing so, the benefits of generative AI can be extended to a sources are vital to enhance the usefulness and applicability of ChatGPT
broader range of students, thereby amplifying its positive impact on the in educational settings. These processes involve cleaning and organizing
global educational landscape (Vilalta-Perdomo et al., 2023). Moreover, data, as well as verifying its relevance and timelines a considerable
the ability of AI technologies like ChatGPT to adapt to various educa­ challenge given the volume and complexity of data produced by
tional contexts and meet the needs of a diverse student body is essential educational institutions (Michel-Villarreal et al., 2023). To ensure the
for their sustained success. Factors such as the "contextuality variable" effective and reliable deployment of AI technologies like ChatGPT in
and the "diversity of users" emphasize the need for developing AI sys­ higher education, institutions must prioritize maintaining high data
tems that are not only robust and secure but also highly adaptable (Díaz, quality and integrity.
2024). This adaptability is crucial to enable the integration of ChatGPT
across different academic disciplines and teaching modalities, from 6. Discussion
traditional classroom settings to fully online courses. For instance,
ChatGPT’s capability to create personalized educational content could In this analysis, we have identified and classified the main and sec­
revolutionize distance learning, providing richer and more accessible ondary topics related to the research questions (RQ2, RQ3, and RQ4) on
learning experiences for students who may be geographically remote or the integration of technologies such as ChatGPT in educational envi­
physically unable to access traditional educational facilities (Zhu et al., ronments. Next, word clouds are presented for each of the research
2023). Furthermore, integrating ChatGPT into higher education requires questions, which visualize the relative prominence of each identified
a proactive effort to ensure that these technologies are used in ways that topic. These images make it possible to clearly distinguish the techno­
respect and uphold the principles of educational equity. It is imperative logical, ethical, and operational challenges that educational institutions
that AI interventions not only achieve technical efficacy but also pro­ face when adopting advanced AI tools. The most prominent topics,
mote an inclusive educational environment that acknowledges and represented in a larger font size, reflect the core concerns of each RQ,
values the cultural and cognitive diversity of the student body while the secondary themes, shown with a smaller font size, comple­
(Vilalta-Perdomo et al., 2023). In conclusion, as educational institutions ment the main discussion. This visual approach facilitates the under­
harness the potential of ChatGPT to enhance learning and teaching, they standing of the key points to consider in the responsible and effective
must also be vigilant in regulating and ethically advancing these tech­ implementation of AI in education.
nologies. Establishing clear regulatory frameworks and engaging in RQ2: Technological Challenges and Integration
continuous research on their social and educational impacts are essen­
tial steps to maximize the advantages and mitigate the risks associated • Major Themes:
with generative AI in academia. ○ Technological challenges: This topic dominates the discussion,

RQ4 What are the main challenges associated with data quality covering aspects such as compatibility, updates, technological
and security in education systems that use AI, and what measures infrastructure, cybersecurity and obsolescence.
can be implemented to ensure data integrity and protection? ○ Data protection and student privacy: This is a crucial issue,

The reliability of AI-based systems remains a critical area of concern. given that data protection and student privacy are fundamental in
Systems like ChatGPT must demonstrate consistent and predictable the integration of new technologies.
performance across different educational settings, free from glitches or • Minor Issues:
errors that could undermine the learning experience or compromise ○ User adaptation and training: This is an important topic, but it is

academic integrity. considered a subtopic within technological challenges.

The technical challenges of implementing and operating AI tech­ ○ Ethical inequalities and algorithmic biases: Although relevant,

nologies such as ChatGPT in higher education are multifaceted. Fig. 6 these issues are secondary in the broader context of technological
outlines data handling must follow rigorous quality and privacy stan­ challenges.
dards. Transparency in data collection and the use of advanced
encryption are essential to avoid risks of exposure of sensitive infor­ RQ3: Personalization and Ethical Challenges
mation, highlighting the critical need to ensure the accuracy and reli­
ability of the data utilized by these technologies (Vilalta-Perdomo et al., • Major Themes:
2023). The effectiveness of an AI model like ChatGPT largely depends on
the quality of the data it is trained on and receives during its operation. A

9
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

○Personalization of learning: This is the central theme of RQ3, methods. However, the ethical implications of personalization must also
focused on how AI can personalize education to suit individual be considered. The reliance on large datasets to drive AI personalization
needs. raises concerns about data privacy and the potential for algorithmic
○ Ethical challenges: Ethical challenges related to data privacy, bias, which could exacerbate existing educational inequities. Recent
equity, and justice in education are central to this question. studies, such as those by Guerrero-Quiñonez et al. (2023), emphasize the
• Minor Issues: importance of developing AI systems that are not only effective but also
○ User diversity and adaptation to different educational con­ ethical and inclusive. This requires a careful balance between leveraging
texts: It is important for personalization, but secondary to ethical AI for personalization and ensuring that all students have equitable
and equity challenges. access to quality education, regardless of their individual circumstances.
○ Impact on educational equity: Although significant, it is mainly Challenge 3: Data Quality and Security
discussed as a consequence of personalization and ethical Ensuring the integrity and security of data in AI-driven educational
challenges. systems is another critical challenge identified in this study. The quality
of the data used by AI technologies like ChatGPT directly impacts their
RQ4: Data Quality and Security effectiveness and reliability. As noted by Vilalta-Perdomo et al. (2023),
data preprocessing and the selection of appropriate data sources are
• Major Themes: essential for maintaining high standards of data quality and integrity.
○ Data quality and integrity: Includes integrity, preprocessing, and Security concerns are equally significant, particularly given the sensi­
data sources, which are critical to the effective functioning of AI tivity of student data. The potential for data breaches or misuse of
technologies. personal information highlights the need for robust security measures,
○ Data security and reliability: This is a crucial issue, as it directly including encryption and regular audits. Michel-Villarreal et al. (2023)
affects the effectiveness and security of the use of AI in education. stress the importance of implementing stringent data protection pro­
• Minor Issues: tocols to safeguard student information and maintain trust in AI tech­
○ Implementation and operation of AI technologies: While nologies. Furthermore, cybersecurity education should be a priority for
important, this topic is more of a means to address data quality and both educators and technical staff to enhance the overall security
security challenges. posture of educational institutions.
Challenge 4: Ethics and Implementation Policies
The discussion section of this study explores the challenges and op­ The ethical implementation of AI in education is a complex and
portunities associated with implementing AI technologies, particularly multifaceted issue that requires the development of clear and compre­
ChatGPT, in higher education. By delving into the intricate dynamics of hensive policies. As AI technologies like ChatGPT become more preva­
technological integration, personalization, data quality, and ethical lent in educational settings, it is crucial to establish ethical frameworks
considerations, this section aims to provide a comprehensive analysis that guide their use. This includes addressing concerns related to data
that connects the findings with broader theoretical perspectives and privacy, algorithmic bias, and the potential for AI to exacerbate existing
recent literature. The discussion is structured around four key challenges inequalities. Díaz (2024) and Holmes et al. (2021) emphasize the need
identified in the study: Technology Integration and Obsolescence, for collaborative efforts between educators, policymakers, and tech­
Personalization and Accessibility, Data Quality and Security, and Ethics nologists to create policies that ensure the responsible and ethical use of
and Implementation Policies. AI in education. These policies should prioritize transparency,
Challenge 1: Technology Integration and Obsolescence accountability, and respect for the privacy and rights of all individuals
One of the most pressing challenges identified in this study is the involved. By establishing such frameworks, educational institutions can
rapid pace of technological obsolescence, which poses significant risks maximize the benefits of AI while minimizing the risks, ultimately
to the sustainability of educational investments in AI. The constant leading to a more equitable and just educational landscape.
evolution of AI technologies like ChatGPT necessitates a dynamic
approach to technology integration, where continuous updates and 6.1. Towards an intelligent educational ecosystem: human-ia synergy,
system scalability are essential. As highlighted in recent literature, collaboration and feedback in the ChatGPT era
modular and scalable systems are critical for managing the rapid
advancement of AI without requiring complete overhauls of existing This paper provides a comprehensive perspective on the challenges
infrastructures (Vilalta-Perdomo et al., 2023). This aligns with Aggarwal and opportunities involved in the implementation of ChatGPT in higher
et al. (2023), who emphasize the need for educational institutions to education, making a significant theoretical contribution to the debate on
adopt flexible architectures that can accommodate ongoing technolog­ the integration of emerging technologies in academic environments.
ical improvements. Moreover, the need for continuous professional Unlike other studies (Alexander et al., 2023), this research focuses on
development for educators cannot be overstated. As new technologies underdeveloped aspects, such as the management of technological
are integrated into educational systems, it is essential that educators are obsolescence and the need for institutional policies that ensure adaptive
equipped with the necessary skills to effectively utilize these tools. This and sustainable integration. In addition, it provides a broader view on
is supported by Abrahão et al. (2021), who argue that user adaptation the personalization of learning, highlighting the importance of
and training are vital for the successful implementation of intelligent balancing the efficiency of automation with ethical principles that
user interfaces in education. Therefore, addressing the challenge of ensure equity (Noroozi et al., 2024; Kasneci et al., 2024).
technological obsolescence requires a multifaceted approach that in­ The analysis conducted through a systematic literature review allows
cludes both technological and human resource development strategies. us to identify the main technological, pedagogical and ethical challenges
Challenge 2: Personalization and Accessibility posed by ChatGPT. Among the most relevant findings, it is highlighted
The ability of AI technologies like ChatGPT to personalize learning that, although AI can enhance the personalization of learning, its use
experiences represents a significant advancement in education. How­ requires clear policies to prevent technological dependence from
ever, this potential also brings with it complex challenges related to compromising academic integrity and exacerbating existing inequalities
accessibility and equity. As discussed by Zhu et al. (2023), personali­ (Zhu et al., 2023). This is especially relevant in diverse educational
zation in education through AI can enhance inclusivity by tailoring contexts, where algorithmic biases can negatively affect marginalized
educational content to meet the diverse needs of students. This is groups if equity audits and transparent algorithms are not implemented
particularly important for students with disabilities or those from (Tlili, 2023).
diverse cultural backgrounds, who may require alternative learning From a theoretical perspective, the study highlights that the adoption

10
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

of ChatGPT in education should not be limited to technical integration, Regarding technological integration and obsolescence, it was iden­
but should foster effective human-IA collaboration. This synergy is key tified that the rapid advancement of educational tools can make them
to maximizing educational benefits by allowing teachers to monitor and obsolete soon after their implementation, which represents a significant
supplement automated responses, ensuring both accuracy and peda­ challenge for institutions. This implies the need for continuous in­
gogical relevance (Lo et al., 2024; Zhu et al., 2024). Ongoing teacher vestments in both technological infrastructure and teacher training. The
training is also essential to adapt to these new tools, thus avoiding skills adoption of modular and scalable architectures appears as an effective
obsolescence and ensuring long-term ethical and efficient use of AI strategy to mitigate this risk, as it facilitates incremental upgrades and
(Adiguzel et al., 2023; Kurban & Şahin, 2024). reduces long-term costs. In this way, technological platforms can adapt
In terms of personalization, ChatGPT’s ability to tailor content ac­ quickly to changes in the educational environment. In addition, ensuring
cording to each learner’s needs represents a significant advantage, but it the constant preparation of personnel is fundamental to maximize the
also poses ethical and logistical challenges related to data privacy. The pedagogical benefits of these tools, ensuring that obsolescence does not
collection and analysis of large volumes of personal data must be negatively affect learning processes. This staff preparation is also closely
managed through secure systems that comply with international regu­ linked to the personalization and accessibility of learning.
lations, ensuring that student information is protected at all times Regarding personalization of learning and accessibility, the use of AI
(Michel-Villarreal et al., 2023; Sharma et al., 2019). In addition, the use has enormous potential to adapt teaching to the individual needs of
of anonymized data and transparency strategies that foster trust in the students, promoting inclusion and facilitating accessibility to diverse
use of these tools are proposed. learning styles. However, this research highlights that personalization
Another key aspect identified is the importance of having modular must be balanced with human intervention to ensure that educational
and scalable infrastructures that allow the continuous updating of processes remain ethical and equitable. AI-generated feedback can be
educational systems without the need for complete renovations. This complemented by peer and teacher input, offering multiple perspectives
approach not only optimizes resources, but also facilitates technological that enhance the development of advanced cognitive skills. In this way,
adaptation, keeping platforms current and functional in a constantly collaboration between teachers and AI is not intended to replace tradi­
evolving environment (Aggarwal et al., 2023; Kurban & Şahin, 2024). tional teaching, but to optimize it by reducing the administrative burden
This study highlights that technological obsolescence is one of the and allowing educators to focus on more meaningful pedagogical as­
biggest challenges for educational institutions, so it is essential to have pects. This combination of automation and human supervision also de­
flexible architectures that support the changing demands of the sector. pends on how educational data is managed.
In addition, it addresses the risk that the massive use of AI in edu­ On data quality and security, it was identified that the management
cation will compromise academic integrity, allowing students to use of large volumes of personal data poses significant risks in terms of
tools such as ChatGPT to generate complete assignments without intel­ privacy and security, which could undermine the trust of students and
lectual effort (Lo, 2023). To mitigate this risk, the use of advanced teachers if not adequately addressed. It is therefore essential that in­
plagiarism detection tools and the promotion of a culture of academic stitutions implement clear data protection policies and ensure trans­
honesty through courses on digital ethics are suggested (Aler Tubella parency in their management. Regular audits and alignment with
et al., 2024). Clear institutional policies and awareness campaigns on international regulations, such as GDPR, are essential to maintain the
the responsible use of AI are essential to ensure integrity in educational integrity of the information collected. Likewise, the use of learning an­
processes. alytics offers a valuable opportunity to optimize educational processes,
This work also proposes that the integration of ChatGPT should be allowing the detection of biases in the results generated by AI and thus
framed within a robust regulatory framework that ensures its fair and ensuring the quality of learning. Responsible data management also lays
accessible use for all students, regardless of their socioeconomic or the foundation for the development of sound ethical frameworks.
cultural background. Institutions should develop policies that regulate In terms of ethics and implementation frameworks, the use of
the use of AI with principles of transparency, accountability, and respect generative AI such as ChatGPT was identified as raising complex ethical
for privacy, encouraging ethical and sustainable adoption (Diaz, 2024; considerations, underscoring the need to formulate clear regulations.
Kasneci et al., 2024). Human intervention is essential to correct for potential algorithmic
Finally, this research brings new perspectives by proposing that the biases that could lead to inequalities between different groups of stu­
implementation of ChatGPT should not be an exclusively technical dents. Institutions should work closely with developers, teachers and
process, but requires a multidimensional approach that integrates regulators to design policies that encourage fair and responsible use of
pedagogical, technological and ethical aspects. The combination of AI. These policies should not only define the limits of automation, but
automation with human intervention, together with a modular infra­ also establish clear roles for human intervention, ensuring that educa­
structure and clear regulatory frameworks, ensures effective and equi­ tional decisions are not left to algorithms alone. The combination of
table integration in higher education. This holistic approach will enable teacher supervision and AI tools is critical to avoid discriminatory out­
institutions to maximize the potential of AI, while managing risks and comes and promote educational inclusion.
fostering an inclusive and sustainable educational environment. Looking ahead, this research highlights the importance of adopting a
proactive approach, based on the continuous updating of policies and
7. Conclusions: towards an ethical and effective implementation the promotion of additional research. It is essential that the proposed
of AI in higher education frameworks be flexible and adaptable to diverse educational contexts,
both in traditional institutions and online learning platforms. Interdis­
The findings of this study comprehensively address the challenges ciplinary collaboration between teachers, technologists and adminis­
and opportunities related to AI integration in higher education. From a trators will maximize the benefits of these tools and minimize the
strategic perspective, clear guidelines are proposed that guide towards associated risks. In addition, the collaborative approach between
safer, more ethical and equitable educational practices. These guidelines humans and AI is applicable at different educational levels, ensuring
highlight critical areas where AI can both transform and complicate that all students, regardless of their background or abilities, have access
educational processes. To ensure effective implementation, the study’s to high-quality learning opportunities. This proactive and collaborative
approach focuses on four key dimensions: technology integration and vision reinforces the importance of a more inclusive and sustainable
obsolescence, learning personalization and accessibility, data manage­ educational future.
ment, and ethics with implementation frameworks, allowing for the This study provides a detailed analysis of the challenges and op­
development of a holistic view that aligns technology, pedagogy, and portunities posed by the implementation of AI as ChatGPT in higher
policy. education. By adopting scalable technology architectures, combining

11
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

human and automated feedback, implementing robust data manage­ Aggarwal, D., Sharma, D., & Saxena, A. B. (2023). Adoption of Artificial Intelligence (AI)
For Development of Smart Education as the Future of a Sustainable Education
ment practices, and formulating clear ethical frameworks, institutions
System. Journal of Artificial Intelligence, Machine Learning and Neural Network.
can effectively integrate these technologies into their educational https://doi.org/10.55529/jaimlnn.36.23.28
programs. AlAfnan, M. A., Dishari, S., Jovic, M., & Lomidze, K. (2023). Chatgpt as an educational
Finally, incorporating the most recent perspectives from the litera­ tool: Opportunities, challenges, and recommendations for communication, business
writing, and composition courses. Journal of Artificial Intelligence and Technology, 3
ture strengthens the theoretical contribution of the study while (2), 60–68.
providing practical guidelines for responsible AI implementation. As Alexander, K., Savvidou, C., & Alexander, C. (2023). Who wrote this essay? Detecting ai-
these technologies evolve, it is essential that educational institutions generated writing in second language education in higher education. Teaching
English with Technology, 23(2), 25–43. https://doi.org/10.56297/BUKA4060/
maintain an active approach to identifying risks and formulating pol­ XHLD5365. Scopus.
icies that maximize benefits. This will ensure a safer, more equitable and Antonio, A.L. (2023). The impact of ChatGPT on college admissions and equity. Stanford
ethical educational future, where technology and pedagogy combine to Graduate School of Education. Retrieved from https://ed.stanford.edu/impact-chat
gpt-college-admissions.
enrich the educational experience. Bird, J., Ekart, A., & Faria, D. (2019). Learning from interaction: An intelligent
networked-based human-bot and bot-bot chatbot system: Contributions presented at
the 18th UK workshop on computational intelligence. https://doi.org/10.1007/9
7.1. Limitations of the study and future work 78-3-319-97982-3_15.
Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the Dangers of
One notable limitation of this study is its reliance on published Stochastic Parrots: Can Language Models Be Too Big?. In Proceedings of the 2021
ACM Conference on Fairness, Accountability, and Transparency (pp. 610–623).
literature, which may not fully capture the most recent advancements in https://doi.org/10.1145/3442188.3445922.
technology due to the inherent delays in scholarly publishing. Further­ Biswas, S. (2023). Role of ChatGPT in computer programming.: ChatGPT in computer
more, the rapidly changing nature of AI technology means that new programming. Mesopotamian Journal of Computer Science, 2023, 8–16.
Buhalis, D., Leung, D., & Lin, M. (2023). Metaverse as a disruptive technology
developments could significantly transform the educational landscape in revolutionising tourism management and marketing. Tourism Management,, 97,
ways that are not yet documented in the current literature. Additionally, Article 104724.
the study primarily focuses on higher education institutions, which may Charan, P.V., Chunduri, H., Anand, P.M., & Shukla, S.K. (2023). From Text to MITRE
Techniques: Exploring the Malicious Use of Large Language Models for Generating
not entirely apply to other educational levels or non-formal learning
Cyber Attack Payloads. arXiv preprint arXiv:2305.15336.
contexts, where dynamics and requirements can vary significantly. Chatterjee, J., & Dethlefs, N. (2023). This new conversational AI model can be your
To address these gaps, it would be beneficial to conduct longitudinal friend, philosopher, and guide. And even your worst enemy. Patterns, 4(1).
studies that monitor the implementation of these technologies over time, Chen, M., Tworek, J., Jun, H., Yuan, Q., Pinto, H.P. de O. et al. (2021). Evaluating large
language models trained on code. arXiv preprint arXiv:2107.03374.
enabling an assessment of their long-term impacts on learning and Chiu, T. K. F., Xia, Q., Zhou, X., Chai, C. S., & Cheng, M. (2023). Systematic literature
educational equity. Moreover, exploring the application of AI in primary review on opportunities, challenges, and future research recommendations of
and secondary education settings could provide valuable insights into artificial intelligence in education. Computers and Education: Artificial Intelligence, 4,
Article 100118. https://doi.org/10.1016/j.caeai.2022.100118
how these tools can be adapted to support educational development Connolly, V. (2023). ChatGPT: Opportunities and challenges for education. Retrieved
throughout different life stages. from https://www.cam.ac.uk/chatgpt-opportunities-and-challenges-for-education.
Damar, M. (2022). What the literature on medicine, nursing, public health, midwifery,
and dentistry reveals: An overview of the rapidly approaching metaverse. Journal of
CRediT authorship contribution statement Metaverse, 2(2), 62–70. https://doi.org/10.57019/jmv.1132962
Dempere, J., Modugu, K., Hesham, A., & Ramasamy, L. K. (2023). The impact of ChatGPT
Iván Miguel García-López: Writing – review & editing, Visualiza­ on higher education. Frontiers in Education, 8. https://doi.org/10.3389/
feduc.2023.1206936
tion, Investigation, Formal analysis. Carina Soledad González Deng, X., & Yu, Z. (2023). A meta-analysis and systematic review of the effect of Chatbot
González: Validation, Data curation. María-Soledad Ramírez-Mon­ technology use in sustainable education. Sustainability, 15(4), 2940. https://doi.org/
toya: Methodology, Conceptualization. José-Martín Molina-Espinosa: 10.3390/su15042940
Diaz, V. (2024). EDUCAUSE. Exploring the opportunities and challenges with generative
Writing – review & editing, Conceptualization. AI Retrieved from: https://er.educause.edu/articles/2024/2/exploring-the-opport
unities-and-challenges-with-generative-ai.
DiGiorgio, A. M., & Ehrenfeld, J. M. (2023). Artificial intelligence in medicine &
Declaration of competing interest ChatGPT: De-tether the physician. Journal of Medical Systems, 47(1), 32.
Dowling, M., & Lucey, B. (2023). ChatGPT for (finance) research: The Bananarama
The authors declare that they have no known competing financial conjecture. Finance Research Letters, 53, Article 103662.
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., et al. (2023).
interests or personal relationships that could have appeared to influence "So what if ChatGPT wrote it?" Multidisciplinary perspectives on opportunities,
the work reported in this paper. challenges and implications of generative conversational AI for research, practice
and policy. International Journal of Information Management, 71, Article 102642.
Firat, M. (2023). What ChatGPT means for universities: Perceptions of scholars and
Acknowledgment students. Journal of Applied Learning and Teaching, 6(1).
Floridi, L., & Chiriatti, M. (2020). GPT-3: Its nature, scope, limits, and consequences.
Minds and Machines, 30(4), 681–694. https://doi.org/10.1007/s11023-020-09548-1
The authors acknowledge the financial and technical support of George, A. S., & George, A. H. (2023). A review of ChatGPT AI’s impact on several
Writing Lab, Institute for the Future of Education, Tecnologico de business sectors. Partners Universal International Innovation Journal, 1(1), 9–23.
Monterrey, Mexico, in the production of this work. Gödde, D., Nöhl, S., Wolf, C., Rupert, Y., Rimkus, L., Ehlers, J., et al. (2023). A SWOT
(strengths, weaknesses, opportunities, and threats) analysis of ChatGPT in the
medical literature: concise review. Journal of Medical Internet Research, 25, e49368.
Funding agency https://doi.org/10.2196/49368
Goralski, M. A., & Tan, T. K. (2022). Artificial intelligence and poverty alleviation:
Emerging innovations and their implications for management education and
The authors acknowledge the financial support of Tecnológico de sustainable development. The International Journal of Management Education, 20(3),
Monterrey through the "Challenge-Based Research Funding Program Article 100662. https://doi.org/10.1016/j.ijme.2022.100662
Guerrero-Quiñonez, A. J., Bedoya-Flores, M. C., Mosquera-Quiñonez, E. F., Mesías-
2022″. Project ID # I003 - IFE001 - C2-T3 – T.
Simisterra, Á. E., & Bautista-Sánchez, J. V. (2023). Artificial Intelligence and its
scope in Latin American higher education. Ibero-American Journal of Education &
References Society Research. https://doi.org/10.56183/iberoeds.v3i1.627
Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Shum, S. B.,
et al. (2021). Ethics of AI in education: towards a community-wide framework.
Abrahão, S., Insfrán, E., Sluÿters, A., & Vanderdonckt, J. (2021). Model-based intelligent
International Journal of Artificial Intelligence in Education, 32, 504–526. https://doi.
user interface adaptation: Challenges and future directions. Software and Systems
org/10.1007/s40593-021-00239-1
Modeling, 20(5), 1335–1349. https://doi.org/10.1007/s10270-021-00909-7
Iftikhar, L. (2023). Docgpt: Impact of chatgpt-3 on health services as a virtual doctor. EC
Adiguzel, T., Kaya, M. H., & Cansu, F. K. (2023). Revolutionizing education with AI:
Paediatrics, 12(1), 45–55.
Exploring the transformative potential of ChatGPT. Contemporary Educational
Technology, 15(3), ep429. https://doi.org/10.30935/cedtech/13152.

12
I.M. García-López et al. International Journal of Educational Research Open 8 (2025) 100401

Ilieva, G., Yankova, T., Klisarova-Belcheva, S., Dimitrov, A., Bratkov, M., & Angelov, D. Pau, A., Fanni, A., Carcangiu, S., Cannas, B., Sias, G., Murari, A., et al. (2019). A machine
(2023). Effects of generative chatbots in higher education. Information (Switzerland), learning approach based on generative topographic mapping for disruption
14(9). https://doi.org/10.3390/info14090492. Scopus. prevention and avoidance at JET. Nuclear Fusion, 59(10). https://doi.org/10.1088/
Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., et al. 1741-4326/ab2ea9
(2023). ChatGPT for good? On opportunities and challenges of large language Pavlik, J. V. (2023). Collaborating with ChatGPT: Considering the implications of
models for education. Learning and Individual Differences, 103, Article 102274. generative artificial intelligence for journalism and media education. Journalism &
Levine, S. (2023). Stanford faculty weigh in on ChatGPT’s shake-up in education. Mass Communication Educator, 78(1), 84–93.
Retrieved from https://ed.stanford.edu/stanford-faculty-chatgpt-education. Rakap, S. (2023). Chatting with GPT: Enhancing individualized education program goal
Kurban, C. F., & Şahin, M. (2024). The Impact of ChatGPT on Higher Education: Exploring development for novice special education teachers. Journal of Special Education
the. AI Revolution. Emerald Publishing Limited. https://doi.org/10.1108/9781 Technology. https://doi.org/10.1177/01626434231211295
837976478. Ren, X., Zhou, P., Meng, X., Huang, X., Wang, Y., Wang, W. et al. (2023). PanGu-
Liebrenz, M., Schleifer, R., Buadze, A., Bhugra, D., & Smith, A. (2023). Generating {\Sigma}: Towards trillion parameter language model with sparse heterogeneous
scholarly content with ChatGPT: Ethical challenges for medical publishing. The computing. arXiv preprint arXiv:2303.10845.
Lancet Digital Health, 5(3), e105–e106. https://doi.org/10.1016/S2589-7500(23) Sallam, M. (2023). ChatGPT utility in healthcare education, research, and practice:
00019-5 systematic review on the promising perspectives and valid concerns. Healthcare,, 11
Lim, W. M., Kumar, S., Verma, S., & Chaturvedi, R. (2022). Alexa, what do we know (6), 887. https://doi.org/10.3390/healthcare11060887
about conversational commerce? Insights from a systematic literature review. Sharma, R., Kawachi, P., & Bozkurt, A. (2019). The Landscape of Artificial Intelligence in
Psychology & Marketing, 39(6), 1129–1155. https://doi.org/10.1002/mar.21654 Open, Online and Distance Education: Promises and Concerns, 14, 1–2.
Lo, C. K. (2023). What Is the Impact of ChatGPT on Education? A Rapid Review of the Stiennon, N., Ouyang, L., Wu, J., Ziegler, D., Lowe, R., Voss, C., et al. (2020). Learning to
Literature. Education Sciences, 13(4), 410. https://doi.org/10.3390/educsci13040410 summarize with human feedback. Advances in Neural Information Processing Systems,
Lo, C. K., Hew, K. F., & Jong, M. S. (2024). The influence of ChatGPT on student 33, 3008–3021.
engagement: A systematic review and future research agenda. Computers & Sun, Y., Wang, S., Li, Y., Feng, S., Chen, X., Zhang, H. et al. (2019). Ernie: Enhanced
Education, 219, Article 105100. https://doi.org/10.1016/j.compedu.2024.105100 representation through knowledge integration. arXiv preprint arXiv:1904.09223.
Lundmark, P. (2022). The real future of the metaverse is not for consumers. Terwiesch, C. (n.d.). Would chat GPT3 get a Wharton MBA.
Michel-Villarreal, R., Vilalta-Perdomo, E., Salinas-Navarro, D. E., Thierry-Aguilera, R., & Tlili, A. (2023). Awareness and acceptance of ChatGPT as a generative conversational AI
Gerardou, F. S. (2023). Challenges and opportunities of generative AI for higher for transforming education by Ghanaian academics: A two-phase study. Journal of
education as explained by ChatGPT. Education Sciences, 13(9). https://doi.org/ Applied Learning & Teaching, 6(2). https://doi.org/10.37074/jalt.2023.6.2.26.
10.3390/educsci13090856 Vargas-Murillo, A. R., Pari-Bedoya, I. N. M. D. L. A., & Guevara-Soto, F. D. J. (2023).
Mijwil, M., Aljanabi, M., & Ali, A. H. (2023). Chatgpt: Exploring the role of cybersecurity Challenges and opportunities of AI-assisted learning: A systematic literature review
in the protection of medical information. Mesopotamian Journal of Cybersecurity, on the impact of ChatGPT usage in higher education. International Journal of
2023, 18–21. Learning, Teaching and Educational Research, 22(7), 122–135. https://doi.org/
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., … Gebru, T. 10.26803/ijlter.22.7.7
(2019). Model Cards for Model Reporting. Proceedings of the Conference on Fairness, Vilalta-Perdomo, E., Salinas-Navarro, D. E., Thierry-Aguilera, R., & Gerardou, F. S.
Accountability, and Transparency, 220–229. https://doi.org/10.1145/328756 (2023). Cross-campus approaches to building a generative AI policy. Education
0.3287596. Sciences, 13(9), 856. https://doi.org/10.3390/educsci13090856
Nadarzynski, T., Bayley, J., Llewellyn, C., Kidsley, S., & Graham, C. A. (2020). Xu, W., & Ouyang, F. (2022). A systematic review of AI role in the educational system
Acceptability of artificial intelligence (AI)-enabled chatbots, video consultations and based on a proposed conceptual framework. Education and Information Technologies,
live webchats as online platforms for sexual health advice. BMJ Sexual & 27(3), 4195–4223. https://doi.org/10.1007/s10639-021-10774-y.
Reproductive Health, 46(3), 210–217. Yeo-Teh, N. S. L., & Tang, B. L. (2023). Letter to editor: NLP systems such as ChatGPT
Noroozi, O., Soleimani, S., Farrokhnia, M., & Banihashem, S. K. (2024). Generative AI in cannot be listed as an author because these cannot fulfill widely adopted authorship
education: pedagogical, theoretical, and methodological perspectives. International criteria. Accountability in Research, 1–3. https://doi.org/10.1080/
Journal of Technology in Education (IJTE), 7(3), 373–385. https://ijte.net/index.ph 08989621.2023.2177160
p/ijte/article/view/845. Zhai, X. (2022). ChatGPT User Experience: Implications for Education. SSRN Electronic
Okey, O. D., Udo, E. U., Rosa, R. L., Rodríguez, D. Z., & Kleinschmidt, J. H. (2023). Journal. https://doi.org/10.2139/ssrn.4312418
Investigating ChatGPT and cybersecurity: A perspective on topic modeling and Zhang, P., & Tur, G. (2024). A systematic review of ChatGPT use in K-12 education.
sentiment analysis. Computers & Security, 135, Article 103476. https://doi.org/ European Journal of Education, 59(2), e12599. https://doi.org/10.1111/ejed.12599
10.1016/j.cose.2023.103476

13