Received: 20 Feb 2024 Revised: 31 May 2024 Accepted: 29 Jun 2024 Published: 30 Jun 2024

Volume 2(1) Pages 107–124 DOI: https://doi.org/10.56279/jicts.v2i1.87

JICTS
Journal of ICT Systems

ChatGPT in Education: Applications, Concerns and

Recommendations
Nicodemus Msafiri Mbwambo1, Peter Benard Kaaya
Department of Information and Communication Technology, Arusha Technical College, Arusha, Tanzania

1
Corresponding author Abstract
Email: [email protected]
Chat Generative Pretrained Transformer (chatGPT) is a language
model that enables users to engage in natural language
Funding information conversations with the system through text. The model is available
This work was not funded.
through an easy-to-use, free web interface with impressive
capabilities, which has contributed to its rapid popularity among
students. This significant interest of students in utilizing chatGPT
underscores the need to understand its impact on education. This
paper presents applications, concerns, and recommendations on the
use of chatGPT in education. The reviewed articles highlight
several educational applications, including assisting students with
writing tasks and helping teachers create learning materials.
However, concerns have been raised about its unregulated use
among students, particularly issues related to academic integrity,
Keywords privacy violations, and potential biases. Regardless, many
AI Models researchers advocate for the integration of chatGPT into the
ChatGpt education system. Some do have reservations and recommend
Deep Neural Networks measures, including learning goals adjustments and development of
Education new assessment forms to embrace ChatGPT in learning. While
Natural Language Processing there is a great potential of chatGPT use in education, further
research is needed to fully understand its implications when used
without appropriate guidance and its long-term effects on students.

1. Introduction

Technology and learning have always been technology has consistently yielded constructive
intertwined for decades. Countless innovations enhancements to global learning endeavors [1].
across the globe have seamlessly integrated into Currently, technology such as tablets and mobile
educational domains to support teaching and phones have widely been used to assist teaching
learning. Audio and video technology as one of the and learning in schools [2]. These technologies
longstanding technologies is a prime example. This collectively profoundly influence enriching

2024 jicts.udsm.ac.tz 107

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

education, transcending geographical barriers to chatGPT that underscores the need to understand its
extend learning even to remote and hazardous impact on education.
environments. To put this growth of chatGPT use into
A standout illustration of technology's perspective, Brandl and Ellis [12] used the amount
transformative potential surfaced during the of time different technologies reached 1 million
unprecedented circumstances of COVID-19 [3]. users after launch. As observed in Figure 1,
Virtual classrooms emerged as a beacon of new chatGPT achieved that growth in just 5 days, 70
educational possibilities, enabling students who days faster than Instagram, which is considered
were unable to physically attend school to continue among the popular social media platforms. We
their learning journeys [4]. Additionally, should point out that chatGPT is not the first
technology holds promise for regions marked by generative AI system, nor is the first to be
elevated levels of violence and social instability, developed by OpenAI. However, chatGPT is one of
presenting a viable solution for fostering education the latest breakthroughs in AI and it is worthwhile
in challenging contexts. to review the current research contextualizing
chatGPT based on the trending frameworks for
In recent years, Artificial Intelligence (AI) has discussing the impact of chatGPT on education.
revolutionized the world in different fields [5],
including teaching and learning. Guan et al. ChatGPT popularity in academia has not
explored over twenty years of Artificial limited its application only on education. There are
Intelligence (AI) and Deep Learning (DL) as many other practical applications [13] (Figure 2).
applied to teaching and learning [6]. Few
applications presented include approaches such as
intelligent tutoring systems, robo-graders,
cognitive assistants [7], and AI assistants [8]. In
November 2022, Chat Generative Pretrained
Transformer (chatGPT) was released. While prior
AI tools are impressive, chatGPT ascends to a
higher echelon by delivering a novel experience,
enabling users to engage in seamless conversations
with a machine using natural language—akin to
conversing with another human being [9]. A user
can ask a range of questions and follow-up
questions. The conversation can start with a simple
question such as What is your name? A fun query
such as "write me a joke" or "write me a movie
scene" to a complex request such as "compose a
conference paper." Because of these features,
chatGPT has rapidly gained a lot of users, ending
up breaking records as the fastest-growing user
base [10], among the users are students [11]. It is Figure 1. Time to reach 1 million users for different
this significant interest of students in utilizing worldwide platforms.

2024 jicts.udsm.ac.tz 108

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

chatGPT does not stand apart. This review intends

to explore its applications, concerns and
researchers’ recommendations on its integration
into the education system.
Literature reviews are necessary as they provide
an in-depth understanding of the existing research
and current progress of a particular topic. This
progress has been slow in some topics, and in
others, has been very fast. For chatGPT and
learning, a huge interest in both the public and
Figure 2. Other applications of ChatGPT.
researchers spiked a lot of studies around the world,
The details regarding these applications are out of which means researchers must keep up with an
the scope of this review, and interested readers are increasing research publications worldwide as new
directed to the following works [13, 14]. insights in the field are almost published every day.
The release of chatGPT ignited a global Studies on chatGPT have been growing
development frenzy in generative AI. Among exponentially every day. Just in April 2023, a
those, at the top is Google chat service called survey by Zhang et al. [25] reported no fewer than
Gemini formally known as Bard [15], which was 500 articles in Google Scholar that included the
officially announced in January 2023. Microsoft word “chatGPT.” At the time we are writing this
came out with Bing Chat [16], and Baidu in China survey, Google Scholar reports 15,500 articles.
created Ernie [17]. Rudolph et al. [18] did a Among those articles, 11,900 include the words
comparison study using 14 multidisciplinary “chatGPT” and “learning”.
questions on chatGPT, GPT-4, Bing Chat, and
The rapid rise in the articles shows researchers’
Bard. The authors concluded from the results that
interest in this topic and underscores the
all bots offer fairly the same performance on the
importance of an up-to-date literature review that
questions. Hence, this review will only focus on
organizes the existing works to help researchers,
ChatGPT. However, most of the findings in this
policymakers, and educators to find results at a
work can be extended or applied to other generative
commonplace. Most recent literature reviews on
AI tools.
chatGPT and education, including the work by
Technology has a lot of contributions to Patricio et al. [26], are limited to articles published
learning, but there are concerns about its use in up to the year 2023. We, therefore, saw a great need
education. For example, research studies conducted for a more updated review in the topic. Other recent
during the 1970s showed various effects of playing reviews include the works by Mai et al. [27] and
video games and McLean expressed early concerns Sumbal et al. [28].
regarding the integration of video games into
This review is expected to explore up to the
learning [19]. Despite those early concerns, video
most recent publications in 2024 to address two
games have continuously been used in training
research questions (RQs):
students, and researchers have continued to explore
their negative [20] and positive [21, 22, 23, 24] RQ 1: What are the applications of chatGPT and the
impacts on education. With the existing knowledge concerns surrounding its integration into
of technologies and their integration into education, education?

2024 jicts.udsm.ac.tz 109

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

RQ 2: Should chatGPT be allowed or limited in also included. We also limited our selection to
education? articles written in English, any articles in other
2. Method languages were excluded.
We chose to conduct a rapid review [29] on the For the articles discussing chatGPT and
topic of chatGPT and learning due to time learning, we focused our selection on publications
constraints, significant public interest in the topic, since 2022. Our choice of more recent studies is
and the rapid evolution in this field. Our search for because chatGPT continues to receive attention,
articles was conducted across four electronic and we believe recent studies provide more
databases: Scopus, Web of Science (WoS), Elton B. understanding of how chatGPT supports or
Stephons Company (EBSCO), and Google Scholar. impedes learning. From the few articles that we
The first three were chosen for their comprehensive selected and reviewed in this group, other highly
coverage, high-quality sources, and robust citation cited articles were identified and obtained from
analysis capabilities. From the search results either Scopus, WoS and EBSCO. Figure 3 shows the
obtained in these three databases, most relevant number of peer reviewed articles selected up to the
articles to our research questions were selected for time this review paper was submitted.
our review. Google Scholar provided us with an
ease of access and a diverse of sources. However, We should point out some of the selected
due to Google Scholar’s wide range of search articles are not peer reviewed publications and were
results, more time was dedicated in assessing the published as news articles, but because of their
results to find high-quality articles that are relevant relevance and contribution to this review, they were
to our topic. Our Google Scholar’s articles selection selected. Table 2 presents the criteria used for
process is shown in Figure 4. The final search of selecting peer reviewed publications from the
articles used in this review was conducted on the databases presented in Table 1.
20th of April 2024 with the following strings used 28
30
as search strings:
# of Papers Published

25
- “Learning and technology” 20
- “Learning and chatGPT” 15 12
Table 1 provides the number of articles selected for 10 7
this review. 5
Table 1. Selected articles on each of the database. 0
Database Number of Articles 2022 2023 2024
Google Scholar 50 Year of Publication
Scopus 18
WoS 08 Figure 3. Number of articles selected in learning and
EBSCO 04 ChatGPT.

A summary of selection criteria used in this Table 3 provides an overview of selected articles
review are provided in Table 2. We divided our with respect to the topics they discuss. We have
search using two main search strings mentioned also included the number of articles that were used
above. Our focus was on conference and journal in explaining key important concepts such as Large
articles mostly relevant to our review. However, Language Models (LLM) and Generative Pre-
some other publications relevant to the topic were Trained Transformer (GPT) as other topics.

2024 jicts.udsm.ac.tz 110

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

Figure 4. Selection of articles through Google Scholar

Table 2. A summary of inclusion and exclusion criteria Table 3: Summary of selected articles for the review.
for articles reviewed.
Topics # of selected articles
Inclusion criteria Exclusion criteria Learning and 23
A result from the Not in the search results Technology
search string “Learning using the string “Learning Learning and ChatGPT 47
and technology” and Technology” Others (LLM, GPT) 10
A result from the Not in the search results
search string “Learning using the string “Learning To respond to our two research questions (RQ
and chatGPT” and chatGPT” 1 and RQ 2), the following steps were followed to
English articles Non-English Articles
perform an in-depth analysis of existing literature,
Academic articles Non-academic articles
Discussing learning and Do not discuss learning organize the selected articles, synthesize the
technology, and and technology and do knowledge gathered and provide a critical
includes its support or not include support or discussion responding to the two research questions
concerns to learning concerns in learning in the introduction:
Discussing learning and Discussing learning and
chatGPT published chatGPT published A. We conducted an in-depth analysis of
since 2022 before 2022 existing literature, selecting articles
Addressing RQ1 and Not addressing RQ1 and relevant to our research questions.
RQ2 RQ2
Peer reviewed articles Non-peer reviewed
articles

2024 jicts.udsm.ac.tz 111

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

B. We synthesized the knowledge gathered Training and Fine-tuning

from the selected articles and organized The training process of the chatGPT model
them accordingly.
begins with unsupervised learning on a vast
C. We provided an overview of chatGPT's
collection of unlabelled data. For instance, GPT-3
special features compared with other
underwent training using over 8 million documents
technologies.
D. We presented the potential applications of and more than 10 billion words [36]. This training
chatGPT in education. is iterative as depicted in step 1 of Figure 5. During
E. We discussed the implications and concerns this phase, the model acquires knowledge of
of allowing the use of chatGPT in patterns and relationships within words, phrases,
education. and sentences.
F. Finally, we discussed both sides of the
argument to answer our research questions, In Step 2, the pre-trained GPT Model from Step
provided researchers recommendations and 1 undergoes fine-tuning using a relatively limited
our opinion in the conclusion section. amount of annotated data through supervised
learning. Figure 5 illustrates this process, where the
3. Overview of chatGPT annotated data consists of selected prompts along
ChatGPT, developed by Open AI [30], is built with corresponding expected output responses
upon its predecessors GPT-1 launched in 2018 [31], generated by human annotators or developers.
GPT-2 [32], GPT-3 [33], and GPT 3.5 [34]. These prompts can either be chosen from OpenAI's
chatGPT is based on LLM implemented as a Deep API requests or created manually. Due to the
Neural Network (DNN) with deep layers of labour-intensive process and costly nature of
transformers, which are specifically engineered to manually generating annotated data, the entire
handle sequential data and have the capability to process becomes time-consuming, yielding
relatively modest outcomes.
produce consistent and human-like responses. To
reach its potential, a chatGPT model is pre-trained
unsupervised on a large corpus of digitized
unlabelled text data (books, articles, and
conversations among other sources) to learn
patterns and relationships in words, phrases, and
sentences [35]. The resulting model is then fine-
tuned through supervised learning for
conversational usage to make it task-specific [31].

The operation principles of chatGPT start with

a user prompt or a question to the model through a
user interface. The model then uses its accumulated
knowledge to generate an appropriate and coherent Figure 5. Training chatGPT model.
response, which is then returned to the user who
The result of step 2 is a Shrink and Fine-Tune
decides from the response to continue the (SFT) Model. This model suffers three primary
conversation or submit another question just like drawbacks due to the limited amount of data used
how two people will hold a conversation. in the fine-tuning process.
2024 jicts.udsm.ac.tz 112
 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

First, the SFT model might output texts that are model at this step is referred to as the PPO model.
very likely not user-attentive, which may result in PPO is an "on-policy" algorithm that uses the trust
responses that are limited in addressing the needs region optimization method to ensure continuous
or query of the user. The second drawback is adaptation of the policy through action from the
misalignment, which causes failure in the SFT PPO model (O) and reward (R) from the Rank
model to capture the intricacies of the target domain Model.
as the model was insufficiently exposed to less data 4. Potential Application of ChatGPT in
during fine-tuning. The third drawback is high Education
scalability costs as scalability requires a model that The capabilities inherent in chatGPT offer
can scale more effectively. To achieve such promising opportunities for innovative integration
scalability, more data must be created for within educational environments. The tool can be a
finetuning. This process is costly. Generally, small powerful teaching assistant or a tool for creativity
data used in fine-tuning a model will have a huge and can play a significant role in enhancing student
impact on its performance, leading to learning.
generalization failure on new examples, and a
struggle to capture more complex queries. Researchers have explored the potential
benefits of integrating chatGPT into educational
These challenges can be mitigated by systems from two primary perspectives: the
expanding the dataset utilized in fine-tuning of student's and the teacher's. In the forthcoming
GPT Model. Nevertheless, employing human sections, we provide a comprehensive summary of
annotators or developers to create a larger dataset various studies conducted within each perspective,
proves to be a laborious and costly endeavor. The shedding light on the diverse insights and findings.
solution to this problem is steps 3 and 4 where
human annotators are used in ranking different 4.1 Students and chatGPT
outputs of the SFT model and create a data set that
is used to train a reward model. ChatGPT emerges as a valuable tool for
students, offering enhanced learning experiences.
In step 3, outputs from the SFT model selected Baidoo-Anu and Ansah [37] claim the tool has
prompts are ranked best to worst by human showcased its strength to craft content ranging from
labelers, creating a new dataset where the rankings a brief paragraph to a detailed research work,
are the labels. Because in this case humans only presenting a convincing or nearly convincing
provide ranks, the process is much faster and can output to earn it several applications in education.
scale up much more effectively. On average, a A public opinion sought through web mining and
dataset is 10 times bigger than demonstration data Natural Language Processing also presents
generated in step 2. chatGPT as a useful writing tool for both students
and educators [38]. The same findings are also
The new data created is used in step 3 to create shared with Mogavi et al. [39] where the authors
a Reward Model (RM), which is used to fine-tune found 78.11% of chatGPT users in their study used
the SFT model generated by step 2 via Proximal it for content creation.
Policy Optimization (PPO).
Kasneci et al. [40] explored the application of
Step 4 uses reinforcement learning, specifically the chatGPT across various educational levels,
PPO algorithm to fine-tune the SFT model by highlighting its positive impact on student learning,
letting it optimize Rank Model. The fine-tuned SFT

2024 jicts.udsm.ac.tz 113

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

including the use of chatGPT for the development - Delivery of real-time feedback and guidance,
of reading and writing skills in elementary school which is necessary for autodidactic learners to
students, learning of language, and writing styles stay on task and achieve their goals on time.
for various subjects for middle and high school - Increased accessibility as chatGPT can be
students, research and writing tasks for university accessed through varieties of devices,
students, facilitate discussions and debates for providing students with more options to
groups and remote learning and combination with access educational materials compared to
speech-to-text or text-to-speech to empower conventional means.
learners with disabilities. Across all levels, - Convenient and flexible learning as chatGPT
chatGPT can provide summaries, and explanations can be accessed anytime and anywhere at
of complex texts, generate practice problems and students' convenience with no appointments.
quizzes, develop students' research and problem- - Self-assessment and reflection as chatGPT
solving skills, and can be used to span the range of offer students a tool they can self-assess and
open research questions. reflect on their progress and learning. The
result of the assessment and reflection can be
In the study conducted by Dijkstra et al. [41], used to pinpoint specific areas that might need
they developed an end-to-end multiple-choice quiz more assistance and direction.
generator. The researchers observed that many of
the generated quizzes demonstrated a high level of Other applications of chatGPT for students
coherence and reasonableness. The authors have been explored by different researchers
emphasize that their work puts chatGPT as a through evaluation of chatGPT's performance on
valuable tool not only for teachers but also for academic exams. One notable area of examination
students, enabling them to enhance their learning is in healthcare education [45]. In this case,
experience by utilizing quizzes to train and evaluate chatGPT was tested on the United States Medical
their knowledge, particularly in the context of exam Licensing Examination [46]. The performance of
preparation. chatGPT in this exam demonstrated its proficiency
in the field, indicating that students in healthcare
Another use of chatGPT is in open education. education can utilize chatGPT as a valuable tool to
The work by Firat explains the potential application enhance their scientific writing skills, promote
of chatGPT in autodidactic students, which research equity, and foster versatility in their
encourages independence and autonomy [42]. The learning. Further, chatGPT can offer essential
study highlights five ways that chatGPT can be assistance to students in personalized learning,
instrumental in supporting self-directed learning: allowing students to concentrate on critical thinking
- Provision of personalized support as chatGPT and problem-based learning.
responds to questions and follow-up
questions, it adjusts the responses according Another academic exam taken by chatGPT was
to students' goals, making the conversation a law school exam. Choi et al. [47] administered the
more tailored to the student. Mhlanga [43] test and chatGPT scored an average of a C+ grade.
and Qadir [44] discuss this support offered by While the analysis of the individual scores on four
chatGPT as a virtual personal tutor for courses taken showed chatGPT could have been a
students. mediocre law student, the authors saw considerable
promise in the use of chatGPT as a tool for
practicing lawyers. From this study, the results

2024 jicts.udsm.ac.tz 114

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

showed a promising tool that could improve the used as an automatic grading system in which
performance of struggling law students or those students' tasks such as research articles, essays, and
students studying under time constraints for their a wide range of written assignments could be
exams. checked for strengths and weaknesses [40].
Undertakings such as creating quizzes and
Many other studies put chatGPT in the spotlight examinations demand a substantial amount of time,
as a tool that can assist students in learning. For and the use of chatGPT can significantly alleviate
example, Wardat et al. [48] reveal the capabilities this time burden for teachers. The time saved can
of chatGPT in providing students with basic be dedicated to other activities such as designing
mathematical knowledge on various topics. Philipp teaching techniques creatively, encouraging
Bitzenbauer [49] introduced two practical independence, and engaging in ways that meet
examples that demonstrate a simple yet an effective students' needs [18].
integration of chatGPT within physics classrooms.
5. Implications and Concerns
These examples specifically aim to nurture critical
thinking skills among secondary school students. 5.1 Academic Honesty and Integrity
4.2 Teachers and chatGPT One major concern in the use of chatGPT by
students is academic honesty and integrity, which
ChatGPT can serve as a valuable tool to
was raised very early by Stokel-Walker [56]. A
augment teachers' efforts in various instructional
recent survey reported about 89% of students in
tasks, enriching the overall learning experience for
American colleges use chatGPT to complete their
students. Nonetheless, it is important to note that
homework assignments, among them, 53% seek
chatGPT should not be considered a replacement
assistance from chatGPT to write articles, 48% use
for educators [50, 51]. In this section, we will
chatGPT during exams, and 22% use it to write
explore how chatGPT is reported to assist teachers
paper outlines [57]. Stanford University presents
in their instructional practices and training
similar concerns on students' use of chatGPT to
methods.
complete different assignments [58]. Unless
One way that chatGPT can assist teachers is students honestly cite their produced work [59, 60],
through the design of learning assessments [52], their submissions are cases of plagiarism, which is
enhancing pedagogical practice, creating becoming hard to detect day by day as chatGPT is
brainstorming ideas and research article outlines, becoming more sophisticated producing responses
and offering a virtual learning environment [53, 54, very close to humans as discussed in [61, 62, 63].
55]. A study by Xiaoming Zhai [51] demonstrated This misuse of chatGPT raises a question of unfair
the potential for chatGPT to assist teachers in learning assessment as judging students on the
developing assessments and learning guidance fairground becomes a challenge.
automatically. The study also shows how chatGPT
Similarly, students who have utilized chatGPT
can become useful by suggesting learning materials
and submit high-quality written assignments are
for learners with special needs to teachers.
more likely to receive better scores than their peers.
Another way chatGPT can be usable by This unfair assessment has far more repercussions
teachers is by assisting them to create quizzes, for students than its immediate benefits including
exams, course syllabus, lesson plans, and other the negative impact on students’ emotions. The
educational resources [37]. ChatGPT can also be same applies for students using chatGPT to

2024 jicts.udsm.ac.tz 115

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

generate essays, who might be unfairly admitted in The integration of ChatGPT into the
colleges, which raises a huge concern as their educational system has sparked significant debate
writings may not reflect their true level of intellect. since it gained widespread attention in 2022.
Researchers, educators, and stakeholders are
It is also a concern that current anti-plagiarism divided in their opinions on the matter. This review
software tools used today are still lagging when it categorizes the analyzed articles into two groups:
comes to chatGPT-generated content [64]. Several those advocating for the adoption of ChatGPT in
GPT detectors are becoming available including education and those recommending its limitations.
OpenAI own’s tool, GPTZero, Content at Scale,
6.1 Embracing chatGPT
Hugging Face, CrossPlag, Writer AI Content
Detector, Draft Goal, and Turnitin, which is also Researchers who agree that chatGPT can be
adding GPT detection capabilities.
effectively integrated into the education system are
5.2 Privacy and Bias Concerns backed by Trust et al. who suggest that the best
approach to combat the misuse of chatGPT is its
The training of chatGPT is fueled by personal
inclusion into the education system and not its
data from the internet without any consent, which
exclusion [70]. To achieve this integration, Trust et
is a privacy violation. Wu et al. discuss four privacy
al. recommend that educators must find the best
risks in chatGPT [65, 66]. This includes privacy
ways to incorporate AI tools such as chatGPT into
leakage due to public data exploitation, privacy
assignments and curricula so students can properly
leakage due to personal input exploitation,
leverage chatGPT to enhance understanding,
emerging new privacy attacks on Large Language
productivity, comprehension, and creativity.
Models and lack of transparency.
Similarly, a study by Zhai [51] suggests an
Like many other AI tools, bias is another adjustment of learning goals that embrace AI tools
concern for chatGPT users. chatGPT has been such as chatGPT as part of the student's learning
trained mostly on internet data, which is not diverse process if chatGPT must be effectively integrated
enough and has bias [67]. To fine-tune chatGPT, into education. Zhai also suggests a new form of
labeled data is used. However, labelers are usually assessment that limits the use of chatGPT and
not representative enough to encompass wide emphasizes critical thinking and creativity in the
viewpoints. Any bias in the data that is used to train follow-up study [65], which also recommends
or fine-tune the model will be reflected in the professional knowledge such as pedagogical
output. The danger in this bias is the unfairness of practice for integrating chatGPT into learning as
chatGPT answers specifically for marginalized necessary knowledge to be acquired by teachers to
groups. This bias is presented in work by Caliskan effectively embrace chatGPT.
et al. [68], which demonstrated an application of
Geo et al. also believes chatGPT can be a
machine learning to ordinary human language
contributing part in education especially after
results in human-like semantic biases. Bias is also
chatGPT proved its ability to write a believable
reported by Bolukbasi et al. [69] to exist in natural
scientific research abstract for high-impact factor
language processing when word embeddings, a
journals with completely generated data. However,
popular framework used to create text data, are
more must be done to maintain rigorous scientific
used.
standards. Geo et al. suggest the inclusion of AI
output detectors in the editorial process and clear
6. Integrating ChatGPT in Education disclosure if these technologies are used [71].

2024 jicts.udsm.ac.tz 116

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

Cutton et al. [61] acknowledges plagiarism to behind these restrictions is the encouragement of
be the main concern for the integration of ChatGPT correct and ethical use of chatGPT, the concerns
in education. However, they suggest that teachers hanging around many academicians were
should provide instructions on the assignments that reinforced by a joint letter signed by well-known
follow a customized submission structure, use figures in tech industries including Elon Musk
rubrics in assessment to ensure students' effort and expressing concerns over AI systems that compete
understanding is fairly assessed, and use manual with human intelligence, and recommended a six-
(in-person) forms of assessment in addition to month suspension on its use and the cessation of
automatic assessment on the assignments to ensure developing large models such as GPT-5 [76].
students' true level of understanding. Cotton et al. 7. Discussions and Recommendations
continue to suggest strategies that can be used to
prevent plagiarism such as educating students on From this review, we can observe that different
plagiarism, receiving a draft of the assignment for researchers have varying perspectives on
review before final submission, using plagiarism chatGPT’s applications and concerns in learning,
detection tools, setting, and communicating clear yet many agree that the rapid proliferation of AI
guidelines on when and how chatGPT can be used, technologies, including chatGPT, will exert a
and finally monitor students work closely. substantial influence on education[77]. In the
6.2 Limiting ChatGPT Use in Education following section, we synthesize and discuss the
review in previous sections and provide
There exists a notable divergence within the recommendations articulated by different
academic discussion regarding the integration of researchers to answer RQ 1 and RQ 2.
chatGPT within educational settings. Some regions
Table 4. Recommendations for embracing of chatGPT
have even adopted a more stringent stance,
in education
imposing restrictions on its use within schools. A
case in point is the decision made by the New York Recommendations for Embracing Articles
City Department of Education (NYC DoE) in ChatGPT
January 2023, wherein they opted to limit the Find best ways to incorporate AI tools [70]
utilization of chatGPT within schools. This such as chatGPT into assignments and
curricula.
decision was motivated by concerns surrounding its
Adjust learning goals that embrace AI [51]
adverse impact on student learning, as well as tools such as chatGPT as part of the
apprehensions regarding the safety and accuracy of student's learning process.
its content, as highlighted by Dan [72]. Develop new forms of assessment that [65]
limits the use of chatGPT and emphasizes
These apprehensions about chatGPT's role in critical thinking and creativity in the
education have spurred numerous educational follow-up study.
Provide professional knowledge such as [65]
institutions worldwide to take precautionary
pedagogical practice for integrating
measures aimed at curbing its usage by students. chatGPT into learning as necessary
Australia has witnessed the prohibition of chatGPT knowledge to teachers.
in various school settings, as noted by Cassidy [73]. Inclusion of AI output detectors in the [71]
Similar measures were taken in Paris’s top editorial process and clear disclosure if [61]
university Science Po [74], RV University in these technologies are used in writing
Bangalore India [75], and many more. While

2024 jicts.udsm.ac.tz 117

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

Table 5. Summary of potential applications and concerns of integrating chatGPT in education.

RQ 1a ChatGPT can be used to: Articles

Facilitate reading, learning languages and writing [40, 39]
Develop students' research and problem-solving skills, span the range of open [40], [45], [46],
research questions, create summaries, and explain complex tasks [37]
Facilitate discussions and debates for groups [40]
Potential applications

Empower learners with disabilities by combining it speech-to-text or text-to-speech [40]

ability
Generate practice problems and quizzes to train and evaluate knowledge [40], [41], [37]
Provide personalized support to students as a virtual personal tutor, real-time [42], [43], [44],
feedback, increased access, self-assessment, and reflection to autodidactic students [46], [53]
Provide cases for practicing lawyers [47]
Provide basic knowledge on various topics [48], [49]
Design learning assessments, teaching methodologies and guidance automatically [51], [53], [52]
[54], [55]
Suggest learning materials for learners with special needs [51]
Grade assignments automatically and classify students by strengths and weaknesses [40], [37]
RQ 1b ChatGPT Articles
Can provide results that may contain factual errors [37], [78], [79]
and Concerns
Implications

Can lead to unfairness in high scores earned by students using chatGPT from low [43], [80]
scores earned by those who do not use chatGPT
Can lead to concerns in academic honesty and integrity in education [38], [56], [57],
[58]
Can violate privacy and can have bias [65], [67]

stakeholders regarding the integration of chatGPT

For the first research question, it is apparent to
in education as presented in Section 6. There is a
us that most researchers see the potential use of
strong advocation from researchers for the careful
chatGPT in education. Both students and teachers
integration of chatGPT into the education system,
can reap the benefits of chatGPT, which is highly
and several studies have demonstrated that
reported to assist in writing tasks. This assistance
embracing chatGPT will yield more benefits than
means less time for teachers in preparation of
restricting its use. These researchers offered several
teaching materials and improved writing skills for
recommendations to such integration as
students. Other potential applications of ChatGPT
summarized in Table 4.
are summarized in Table 5. While the benefits are
clear for the students, concerns raised by During our review, we found that relatively few
researchers are based on chatGPT’s uncontrolled articles provided findings supporting the restriction
of chatGPT in education. However, we identified
use in learning, which may lead to academic
different measures taken by governments in
dishonesty. These concerns were reviewed in
different countries and universities to stop or
Section 5 and a summary is provided in Table 5.
significantly limit the use of chatGPT by students
For the second question, our review uncovers until a deeper understanding of its capabilities and
diverse of opinion from different researchers and impact on education is achieved.

2024 jicts.udsm.ac.tz 118

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

Therefore, it is made clear that chatGPT has a Considering the applications and concerns
lot of applications in education but its uncontrolled discussed, we believe that ChatGPT has the
use by students is raising concerns. Both its potential to revolutionize education for both
applications and concerns are summarized in Table students and teachers. However, its use must be
5 thus responding to RQ 1. carefully controlled until all concerns are
thoroughly addressed.
As for RQ 2, many researchers support the
integration of chatGPT into our education system In the meantime, we recommend more research
and recommended how to do it as summarized in efforts to address ethical considerations
Table 4. Other education stakeholders felt that more surrounding the use of chatGPT in education,
work was needed to address all concerns relating to investigate how it can be integrated into
the use of chatGPT in education and opted a collaborative learning environments, explore the
limitation or a complete ban in the use of chatGPT long-term impact of chatGPT integration on
in education. students’ academic achievement, learning
outcomes, and critical thinking skills.
8. Conclusion

CONTRIBUTIONS OF CO-AUTHORS
Nicodemus Msafiri Mbwambo [ORCID: 0009-0008-2616-0607] Conceived the idea and wrote the paper
Peter Benard Kaaya [ORCID: 0009-0001-9846-4211] wrote the paper

2024 jicts.udsm.ac.tz 119

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

REFERENCES

[1] Garrison, D. R., An Analysis and evaluation of audio teleconferencing to facilitate education at a
distance, American journal of distance education, 4(3): p.13–24, 1990.
[2] Foti, M. K., Mendez, J., Mobile Learning: How students use mobile devices to support learning,
Journal of literacy and technology, 15(3): p. 58-78, 2014.
[3] Pandey, D., Ogunmola, G. A., Enbeyle, W., Abdullahi, M., Pandey, B. K., Pramanik, S., COVID-19:
A framework for effective delivering of online classes during lockdown, Human Arenas, 5(2): p. 1–5,
2021.
[4] Asmara, R., Teaching english in a virtual classroom using whatsapp during COVID-19 pandemic,
Language and education journal, 5(1): p. 16–27, 2020.
[5] Xu, Y., Liu, X., Cao, X., Huang, C., Liu, E., Qian, S., Liu, X., Wu, Y., Dong, F., Xiu, C., Qiu, J., Hua,
K., Su, W., Wu, J., Xu, H., Han, Y., Fu, C., Yin, Z., Liu, M., Roepman, R., Zhang, J., Artificial
Intelligence: a powerful paradigm for scientific research, The Innovation, 2(4): p. 100–179, 2021.
[6] Guan, C., Mou, J., Jiang, Z., Artificial Intelligence innovation in education: A Twenty-Year data-driven
historical analysis, International journal of innovation studies, 4(4): p. 134–147, 2020.
[7] Daley, S., Pennington, J., Alexa the Teacher’s Pet? A review of research on virtual assistants in
education, In EdMedia + innovate learning, waynesville, Association for the advancement of
computing in education (AACE), p. 138–146, 2020.
[8] Kim, J., Merrill, K., Xu, K., Sellnow, D. D, My Teacher is a machine: understanding students’
perceptions of AI teaching assistants in online education, International journal of human-computer
interaction, 36(20): p. 1902–11, 2020.
[9] Kirmani, A. R., Artificial Intelligence-enabled science poetry, ACS Energy Letters, 8(1): p. 574–76,
2023.
[10] Krystal, H., ChatGPT sets record for fastest-growing user base - analyst note, Reuters, Link:
https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-
2023-02-01/, 2023,. Accessed 30 March 2024.
[11] Castillo, A. G. R., Rivera, H. V. H., Teves, R. M. V., Lopez, H. R. P., Reyes, G. Y., Rodriguez, M. A.
M., & Arias-Gonzáles, J. L. Effect of Chat GPT on the digitized learning process of University
students, Journal of namibian studies: History politics culture, 33. p. 1–15, 2023.
[12] Brandl, R., Ellis, C., ChatGPT Statistics and User Numbers 2023 - OpenAI Chatbot, tooltester, Link:
https://www.tooltester.com/en/blog/chatGPT-statistics/, 2024. Accessed 30 March 2024.
[13] Nazir. A., Wang, Z., A comprehensive survey of ChatGPT: Advancements, applications, prospects,
and challenges, Meta-Radiology, 1(2), 2023.
[14] Abdullah, M., Madain, A., Jararwah, Y., ChatGPT: Fundamentals, Applications and Social Impacts,
In 2022 Ninth International Conference on Social Networks Analysis, Management and Security
(SNAMS), p. 1-8, 2022.
[15] WIKIPEDIA, Gemini (chatbot), Wikipedia, Link: https://en.wikipedia.org/wiki/Gemini_(chatbot).
Accessed 30 March 2024.
[16] Mehdi, Y., Reinventing search with a new AI-powered microsoft bing and edge, your copilot for the
web, Microsoft, 2023. [Online]. Accessed 30 March 2024.

2024 jicts.udsm.ac.tz 120

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

[17] David, E., Baidu launches Ernie chatbot after Chinese government approval, The verge. Link:
https://www.theverge.com/2023/8/31/23853878/baidu-launch-ernie-ai-chatbot-china. Accessed 30
March 2024.
[18] Rudolph, J., Tan, S., Tan, S., War of the chatbots: Bard, Bing Chat, chatGPT, Ernie and beyond. the
new AI gold rush and its impact on higher education, Journal of applied learning and teaching, 6(1):
p. 364-389, 2023.
[19] McLean, H. W., Are simulations and games really legitimate?, Audiovisio Instruction, 7, p. 12-13,
1978.
[20] Jordan, A. B., Romer, D., Media and the well-being of children and adolescents, Oxford University
Press, 2014.
[21] Griffiths, M. D., The educational benefits of videogames, Education and health, 20(3):p. 47-51, 2002.
[22] Annetta, L. A., Video Games in Education: Why They should be used and how they are being used,
Theory into practice, 47(3):p. 229–39, 2008.
[23] Chuang, T. Y., Chen, W, F., Effect of computer-based video games on children: an experimental study,
IEEE international workshop on digital game and intelligent toy enhanced learning, p. 114–18, 2007.
[24] Halbrook, Y, J., O’Donnell, A. T., Msetfi, R. M., When and how video games can be good: a review
of the positive effects of video games on well-being, Perspectives on psychological science, 14(6): p.
1096–1044, 2019.
[25] Zhang, C., Zhang, C., Li, C., Qiao, Y., Zheng, S., Dam, S. K., Zhang, M., Kim, J. U., Kim, S. T., Choi,
J., Park, G., Bae, S., Lee, L., Hui, P., Kweon, I. S., Hong, C. S., One small step for generative AI, one
giant leap for AGI: a complete survey on chatGPT in AIGC Era, arXiv, 2023.
[26] Patrício, M. R., Gonçalves, B. F., ChatGPT: Systematic review of potentials and limitations in
education, Information technology and systems. ICITS 2024. Lecture Notes in Networks and systems,
933, p. 339–348, 2024.
[27] Mai, D. T. T., Da, V. C., Hanh, N. V., The use of ChatGPT in teaching and learning: a systematic
review through SWOT analysis approach, Frontiers in education, 9, p. 1328769, 2024.
[28] Sumbal, A., Sumbal, R., Amir, A., Can ChatGPT-3.5 pass a medical exam? A systematic review of
ChatGPT's performance in academic testing, Journal of medical education and curricular
development, 11, 2024.
[29] Khagura, S. D., Konnyu, K. J., Grimshaw, J., Cushman, R., Moher, D., Evidence summaries: The
evolution of a rapid review approach, Systematic Review, 1(1): p. 10, 2012.
[30] ChatGPT, OpenAI. Link: https://openai.com/chatgpt. Accessed 31 March 2024.
[31] Radford, A., Narasimhan, K., Salimans, T., Sutskever, I., Improving language understanding by
generative pre-training, Open AI, 2018.
[32] Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., Sutskever, I., Language models are unsupervised
multitask learners, OpenAIblog, 1(8): p. 9, 2019.
[33] Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J, D., Dhariwal, P., Neelakantan, A., Shyam,
P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R.,
Ramesh, A., Language models are few-shot learners, Advances in Neural Information Processing
Systems, 33, p. 1877–1901, 2020.
[34] Dale, R., GPT-3: What’s It Good for?, Natural Language engineering, 27(1): p. 113–18, 2021.
[35] Badzianowski, P., Vulić, I., Hello, It's GPT-2 -- How Can I Help You? towards the use of pretrained
language models for task-oriented dialogue systems, arXiv, 2019.
[36] Kublik, S., Saboo, S., GPT-3: Building Innovative NLP products using large language models,
O'Reilly Media, 2022.

2024 jicts.udsm.ac.tz 121

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

[37] Baidoo-Anu, D., Ansah, L. O., Education in the era of Generative Artificial Intelligence (AI):
understanding the potential benefits of chatGPT in promoting teaching and learning, Journal of AI,
7(1): p. 52–62, 2023.
[38] Rajeb, A., Rajeb, K., Appolloni, A., Treiblmaier, H., Iranmanesh, M., Exploring the impact of
ChatGPT on education: A web mining and machine learning approach, International Journal of
Management Education, 22(1): p. 100932, 2024.
[39] Mogavi, R. H., Deng, C., Kim, J. J., Zhou, P., Kwon, Y. D., Metwally, A. H. S., Tlili, A., Bassanelli,
S., Bucchiarone, A., Gujar, S., Nacke, L. E., and Hui, P., ChatGPT in education: A blessing or a curse?
A qualitative study exploring early adopters’ utilization and perceptions, Computers in Human
Behavior: Artificial Humans, 2(1), 2024.
[40] Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh,
G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., ChatGPT for
good? on opportunities and challenges of large language models for education, Learning and
individual differences, 103, p. 102274, 2023.
[41] Dijkstra, R., Genc, Z., Kayal, S., and Kamps, J., Reading comprehension quiz generation using
generative pre-trained transformers, Fourth international workshop on intelligent textbooks 2022 co-
located with 23d International Conference on Artificial Intelligence in education, 3192, p. 4-17, 2022.
[42] Firat, M., How chat GPT can transform autodidactic experiences and open education, Department of
Distance Education, Open Education Faculty, 2023.
[43] Mhlanga, D., Open AI in Education, the Responsible and Ethical Use of ChatGPT Towards Lifelong
Learning, In FinTech and Artificial Intelligence for sustainable development. sustainable development
goals series, Palgrave Macmillan, p. 387–409, 2023.
[44] Qadir, J., Engineering education in the era of chatGPT: promise and pitfalls of generative AI for
education, 2023 IEEE Global Engineering Education Conference (EDUCON), p. 1-9, 2023.
[45] Sallam, M., ChatGPT Utility in Healthcare education, research, and practice: systematic review on
the Promising Perspectives and Valid Concerns, Healthcare, 11(6): p. 887, 2023.
[46] Gilson, A., Safranek, C. W., Huang, T., Scrates, V., Chi, L., Taylor, R. A., Chartash, D., How Does
chatGPT perform on the United States medical licensing examination? the implications of large
language models for medical education and knowledge assessment, JMIR Medical education, 9(1): p.
e45312, 2023.
[47] Choi, J. H., Hickman, K. E., Monahan, A., Schwarcz, D., ChatGPT goes to law school, J. Legal educ.,
71, p. 387, 2021.
[48] Wardat, Y., Tashtoush, M. A., Alali, R., Jarrah, A, M., ChatGPT: A Revolutionary Tool for Teaching
and Learning Mathematics, Eurasia journal of mathematics, science and technology education, 19(7):
p. 1–18, 2023.
[49] Bitzenbauer, P., ChatGPT in Physics Education: A Pilot Study on Easy-to-Implement Activities,
Contemporary Educational Technology, 15(3): p. 430, 2023.
[50] Ausat, A. M. A., Massang, B., Efendi, M., Nofirman., Riady, Y., Can Chat GPT Replace the Role of
the Teacher in the Classroom: A Fundamental Analysis, Journal on Education, 5(4): p. 16100–06106,
2023.
[51] Zhai, X., ChatGPT for Next Generation Science Learning, XRDS: Crossroads, The ACM Magazine
for Students, 29(3): p. 42-46, 2023.
[52] Parker, L., Sokkar, A., Karakas, A., Carter, C., Graduate instructors navigating the AI frontier: The
role of ChatGPT in higher education, Computers and education open, 6, p. 100166, 2024.
[53] Sok, S., Heng, K., ChatGPT for Education and Research: A Review of Benefits and Risks, Cambodian
journal of educational research, 3(1): p. 110–121, 2023.

2024 jicts.udsm.ac.tz 122

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

[54] Divito, C. B., Katchikian, B. M., Gruenwald, J. E., Burgoon, J. M., The tools of the future are the
challenges of today: The use of ChatGPT in problem-based learning medical education, Medical
teacher, 43(3): p. 320–322, 2024.
[55] Al Shloul, T., Mazhar, T., Abbas, Q., and Iqbal, M., Yaseen Ghadi, Y., Malik, F., Hamam, H., Role of
activity-based learning and ChatGPT on students' performance in education, Computers and
education: artificial intelligence, 6, p. 100219, 2024.
[56] Stokel-Walker, C., AI Bot chatGPT Writes Smart Essays — Should Professors Worry?, Nature, 2022.
[57] Tangermann, V., 89 Percent of college students admit to using chatGPT for homework, study claims,
futurism. Link: https://futurism.com/the-byte/students-admit-chatgpt-homework. Accessed 31 March
2023.
[58] Cu, M. A., Hochman, S., Scores of Stanford students used chatGPT on final exams, The Stanford
Daily, Link: https://stanforddaily.com/2023/01/22/scores-of-stanford-students-used-chatgpt-on-final-
exams-survey-suggests/. Accessed 31 March 2024.
[59] Kleebayoon, A., Wiwanitkit, V., Artificial Intelligence, Chatbots, Plagiarism and Basic Honesty:
Comment, Cellular and Molecular Bioengineering, 16(2): p. 173–174, 2023.
[60] Thurzo, A., Strunga, M., Urban, R., Surovková, J., Afrashtehfar, K. I., Impact of artificial intelligence
on dental education: a review and guide for curriculum update, Education sciences, 13(2): p. 150,
2023.
[61] Cotton, D. R. E., Cotton, P. A., Shipway, J. R., Chatting and cheating: ensuring academic integrity in
the era of chatGPT, Innovations in education and teaching international, 61(2): p. 228–239, 2023.
[62] Else, H., Abstracts written by chatGPT fool scientists, Nature, 613(7944): p. 423, 2023.
[63] Shiri, A., ChatGPT and Academic Integrity, Information Matters, 3(2), 2023.
[64] Jemmy., Mia, A., Wahdah, Joshua, W., Sabri., Impact of GPT Chat in Higher Education Learning,
Journal International of Lingua and Technology, 3(1): p. 43-57, 2024.
[65] Wu, X., Duan, R., Ni, J., Unveiling security, privacy, and ethical concerns of chatGPT, Journal of
information and intelligence, 2(2): p. 102-115, 2023.
[66] Lo, C, K., What is the impact of chatGPT on education? a rapid review of the literature, Education
sciences, 13(4): p. 410, 2023.
[67] Bender, E. M., Gebru, T., McMillan-Major, A., Schmitchell, S., On the dangers of stochastic parrots:
can language models be too big?, ACM Conference on Fairness, accountability, and transparency, p.
610–623, 2021.
[68] Caliskan, A., Bryson, J. J., Narayanan, A., Semantics derived automatically from language corpora
contain human-like biases, Science, 356(6334): p. 183–186, 2017.
[69] Bolukbasi, T., Chang, K., Zou, J., Saligrama, V., Kalai, A., Man is to computer programmer as woman
is to homemaker? debiasing word embeddings, 30th International conference on neural information
processing systems, 29, p. 4356–4364, 2016.
[70] Trust, T., Whalen, J., Mouza, C., ChatGPT: Challenges, Opportunities, and implications for teacher
education, Contemporary issues in technology and teacher education, 23(1): p. 1–23, 2023.
[71] Gao, C. A., Howard, F. M., Marvov, N., Dyer, E. C., Comparing scientific abstracts generated by
chatGPT to original abstracts using an artificial intelligence output detector, plagiarism detector, and
blinded human reviewers, bioRxiv, 2022.
[72] Rosenzweig-Ziff, D., New York City blocks use of the chatGPT bot in its schools, The Washington
Post. Accessed 31 March 2024.
[73] Cassidy, C., Queensland Public Schools to Join NSW in banning students from chatGPT, The guardian.
Accessed 31 March 2024.

2024 jicts.udsm.ac.tz 123

 JICTS
Mbwambo et al. Volume 2(1) Pages 107–124

[74] Reuters., French University Bans Use of chatGPT to Prevent Plagiarism., Reuters, [Online].
[Accessed 31 March 2024], 2023.
[75] Chakravarti, A., Bengaluru University Bans chatGPT after Students Found Using AI Tool for Exams,
Assignments, and Lab Tests, India Today. Accessed 31 March 2024.
[76] Samuel, J., Response to the March 2023 Pause Giant AI Experiments: an open letter by Yoshua Bengio,
Signed by Stuart Russell, Elon Musk, Steve Wozniak, Yuval Noah Harari, and Others, Social Science
Research Network, 2023.
[77] Eke D. O., ChatGPT and the Rise of Generative AI: Threat to Academic Integrity?, Journal of
Responsible Technology, 13, p. 100060, 2023.
[78] Gordijn, B., Have, H, T., ChatGPT: Evolution or revolution?, Medicine, Health Care and Philosophy,
26(1): p. 1-2, 2023.
[79] Van Dis, E, A., Bollen, J., Zuidema, W., Rooiji, R. V., & Bockting, C. L., ChatGPT: five priorities
for research, Nature, 614 (7947):p. 224–226, 2023.
[80] Taecharungroj, V., What Can chatGPT Do? Analyzing early reactions to the innovative AI chatbot on
Twitter, Big Data and Cognitive Computing, 7(1): p. 35, 2023.

2024 jicts.udsm.ac.tz 124