2023 5th International Workshop on Artificial Intelligence and Education (WAIE)

Analyzing elementary teachers' perception of

ChatGPT
Jiyong Moon Seong Baeg Kim
Computer Education Major, Graduate School Computer Education Major, Graduate school
2023 5th International Workshop on Artificial Intelligence and Education (WAIE) | 979-8-3503-0706-1/23/$31.00 ©2023 IEEE | DOI: 10.1109/WAIE60568.2023.00015

Jeju National University Jeju National University

Jeju, Republic of Korea Jeju, Republic of Korea
[email protected] [email protected]

Abstract— As the public's interest in ChatGPT has increased, and personalized instruction in the field of education. Hence,
research on its educational use has also been actively conducted. various studies are currently being conducted from an
However, understanding the perception of individual teachers educational perspective.
regarding ChatGPT is incomplete. Therefore, in this study, we
investigated teachers' perceptions according to their profiles. However, most educational research on ChatGPT needs to
After reviewing previous studies on teachers' perceptions of be vetted to ensure that it reflects the opinions of educators [5].
ChatGPT, we gathered expert opinions to refine the survey There is no dispute over the fact that the role of educators is
questions. Based on the results of this study, we aim to explore how extremely crucial as key agents in education[6]. Nevertheless,
elementary school teachers can develop positive perceptions of while there are comprehensive studies that examine how
ChatGPT in school considering their profiles. teachers perceive ChatGPT, there is still a lack of research on
how their perceptions of ChatGPT vary depending on their
Keywords—ChatGPT, teacher’s perception, educational characteristics and how educational strategies for utilizing
application, elementary school ChatGPT should be tailored accordingly.
II. REVIEW OF RELATED RESEARCH
I. INTRODUCTION A. Perception literature
Early applications of artificial intelligence (AI) technology To analyze the literature, we examined teachers' perceptions
were primarily focused on solving specific problems within of ChatGPT, but there was not enough data available. Therefore,
limited scopes and contexts. However, recent advancements we analyzed studies that explored teachers' perceptions of
have led to the forms that closely mimic human neural networks ChatGPT or AI within the context of ChatGPT being embedded
and intelligence[1,2]. A prime example of this is ChatGPT in AI.
(Chat-Generative Pre-trained Transformer). Launched on
November 30, 2022, ChatGPT demonstrated a level of Miyoung Ryu analyzed that elementary school teachers
responsiveness that was radically different from any chatbot AI expected AI to bring major changes to our daily lives.
service we had seen up to that point, gaining over 100 million Particularly, she noted that experiences, training, and research
monthly users within two months of launch [3]. Based on the schools related to software education can be a positive factor.
principle of large-scale language model (LLM), ChatGPT Therefore, it is important to provide extensive training on AI in
provided a simple conversational AI service that can be accessed education[6].
with a single click Soo-young Lee found that elementary school teachers
People who experienced the new AI service reacted positively evaluated AI technology for providing individualized
enthusiastically, as advances in AI technology are expected to customized content based on learners' interests and
have a significant impact on future economies, societies, characteristics. However, their understanding of AI technology
cultures, and education. This anticipation has motivated the was negative. In addition, although beginners' interest in AI
general public to explore way to leverage AI technology [4]. education was high, it was not as high as that of active adopters.
Additionally, ChatGPT offers the potential to enhance learning Therefore, there is a need to improve understanding of AI
technology and make AI education more accessible[7].
This research was supported by the MSIT(Ministry of Science and ICT), Korea,
under the National Program for Excellence in SW)(2018-0-01863) supervised by
the IITP(Institute of Information & communications Technology Planning &
Evaluation) in 2023.

979-8-3503-0706-1/23/$31.00 ©2023 IEEE 43

DOI 10.1109/WAIE60568.2023.00015
Authorized licensed use limited to: Zhejiang University. Downloaded on May 01,2024 at 16:06:02 UTC from IEEE Xplore. Restrictions apply.
 Keun-Jae Kim and his colleagues analyzed that the use of AI It has been validated and extended by many researchers, leading
in education is suitable for supporting activities during in-class to the development of enhanced technology acceptance models
hours. They emphasized that in order to actively utilize AI, it is such as TAM2 and TAM3.
necessary to build a learning environment that can utilize AI
effectively [8]. In TAM2, factors comprising perceived usefulness were
presented as social norms, image, and voluntariness, which are
Han-byeol Kim analyzed the perceptions of middle school related to the social influence process, while factors related to
teachers and argued that it is essential to expand AI basic literacy the cognitive instrumental process, including task relevance,
education and develop customized AI education and curriculum result quality, result demonstrability, and perceived ease of use,
based on middle school curriculum. He emphasized the were introduced to predict user behavior[6].
necessity of implementing teacher research groups and a system
for leading teachers to facilitate the spread of AI education in TAM3 introduced four factors: computer self-efficacy,
classrooms[9]. computer anxiety, computer enjoyment, and perceived external
support, and proposed that perceived ease of use would be
Sumin Hong comprehensively analyzed the perceptions of moderated by subsequent perceived enjoyment and objective
elementary school teachers regarding the educational utilization usability[7].
of ChatGPT. The result of the study showed that although there
was some experience with using ChatGPT, there was limited
educational application experience. Additionally, there was a
negative view towards ethical concerns. The study identified the
need to develop a customized educational program for
elementary school teachers[5].
After analyzing the relevant studies, we found that while
there are many studies on teachers' perceptions of AI, there are
relatively few studies focusing on teachers' perceptions of
ChatGPT. Furthermore, studies on teachers' perceptions of
ChatGPT are in their early stages, and many of them are broadly
based on teachers as a whole, so there is a lack of detailed
insights into teachers' perceptions of ChatGPT based on their
individual characteristics.
B. ChatGPT
The ChatGPT is a natural language processing model based
on GPT-3 developed by OpenAI. It is a transformer model fine-
tuned through supervised and reinforcement learning, capable of
various natural language processing tasks such as text
generation, translation, summarization, and question answering.
As a result, ChatGPT engages in conversations using natural
language like humans. Unlike previous rule-based chatbots, it
generates language and comprehends conversation context,
demonstrating its interactive capabilities[4].
C. TAM
TAM is a model designed to predict computer user
acceptance, explaining that perceived usefulness and ease of use
influence attitudes towards technology use, which in turn impact Fig 1. Technology Acceptance Model 3 (TAM3)
technology acceptance intentions and actual adoption. These
perceived usefulness and ease of use can be influenced by III. RESEARCH METHOD
external factors and, furthermore, perceived ease of use affects
perceived usefulness. The technology acceptance model has A. Analysis Plan
demonstrated excellent performance in predicting the use and To understand elementary teachers' perceptions of ChatGPT
adoption of information systems[5]. based on their characteristics, we categorized them into gender,
years of experience, specific university major, hours of SW·AI
training in the past two years, and SW·AI educational
background. To achieve the research objectives, a survey
targeting elementary teachers will be conducted over a month.
We plan to analyze responses from over 100 elementary teachers,
ensuring a balanced distribution of gender and years of teaching
experience to enhance fairness in survey results. The
classification of teachers’ subject characteristics is summarized
in Table 1.
Fig. 1. Technology Acceptance Model (TAM)

44

Authorized licensed use limited to: Zhejiang University. Downloaded on May 01,2024 at 16:06:02 UTC from IEEE Xplore. Restrictions apply.
 TABLE I. CLASSIFICATION OF TEACHERS’ PROFILES TABLE II. EDUCATIONAL UTILITY QUESTIONNAIRE
Section Question
Category Item What do you consider to be the closest thing to AI
(ChatGPT)?
Male An artificial entity capable of human-like interaction
Gender Robots that assist in performing intellectual tasks
Female Computer systems capable of human-like thinking or
Less than 5 years reasoning
Concept of AI
Computer programs that swiftly analyze or calculate
More than 5 years but less than 10 years (ChatGPT)
various pieces of information
Career A visible form of processing diverse information and
More than 10 years but less than 15 years
data through computer programs
More than 15 years but less than 20 years An intelligent system integrating knowledge from
various fields such as engineering, natural sciences,
More than 20 years
education, and more
Korean AI (ChatGPT) is a convenient technology.
Positive
AI (ChatGPT) will bring happiness to people.
Mathematics perceptions of
AI (ChatGPT) will create more jobs than it currently
Social Studies AI (ChatGPT) does.
Negative AI (ChatGPT) is a difficult-to-control technology.
Science
perceptions of AI (ChatGPT) will pose a threat to people.
English AI (ChatGPT) AI (ChatGPT) will lead to a reduction in people's jobs.
Undergradu Which subjects are suitable for using ChatGPT in
Ethics
ate major teaching, learning, and assessment methods?
Practical Art Korean Practical Art
Computer Mathematics Computer
Relevance of Social Studies Music
Music Science Art
AI (ChatGPT)
Art to subjects English Education
and grade Ethics Creative experience activities
Education
levels From which grade level is it appropriate to start using
None ChatGPT?
1st to 2nd grade
More than 1 hour but less than 30 hours 3rd to 4th grade
Hours of
SW·AI 5th to 6th grade
What do you think is the closest to AI (ChatGPT)
training in More than 31 hours but less than 60 hours
education?
the past two
years Education on developing artificial intelligence
More than 61 hours but less than 90 hours Concept of AI
Education on utilizing artificial intelligence in life
(ChatGPT)
Education on integrating artificial intelligence into
More than 91 hours education
curriculum
Low educational competency Education on understanding the concepts and
principles of artificial intelligence
SW·AI Educational competency is on the lower side. I am interested in teaching the principles and key
educational concepts of AI (ChatGPT)
competency Educational competency is on the higher side. I am interested in addressing various ethical issues
AI (ChatGPT) related to AI (ChatGPT).
Educational competency is high. interest in
I am interested in utilizing AI (ChatGPTP as a teaching
teaching
and learning tool.
I am interested in using AI(ChatGPT) as an assessment
B. Educational Utility Evaluation Tool tool.
Due to a lack of research on the perception of ChatGPT I can explain the principles and key concepts of AI
educational applications, we conducted a thorough analysis of Teaching (ChatGPT).
competency I can address various ethical issues related to AI
previous studies, including teachers' perceptions of AI, to select AI (ChatGPT) (ChatGPT).
an appropriate assessment tool. After reviewing various education I can use AI as a teaching and learning tool.
previous research, Han-byeol Kim’s assessment tool was I can utilize AI as an assessment tool.
adopted as it met the purpose of this study. However, since Students should be taught the principles and key
Kim's tool focuses on secondary school teachers' perceptions of concepts of AI (ChatGPT).
AI and AI education, the term "AI" in the original tool was They should be enabled to address ethical issues
changed to focus on ChatGPT. In addition, unlike middle school Perception of related to AI (ChatGPT).
need for AI Students should be instructed on how to develop AI
teachers, elementary school teachers instruct all subjects, and (ChatGPT) (ChatGPT).
their grade range is broader. Therefore, one section (AI and education They should also be taught how to utilize AI
subject and grade relevance) was consulted by two experts. The (ChatGPT) as a learning tool.
specific questions of the survey is presented in Table 2. Teacher training on AI (ChatGPT) education is
needed.
AI (ChatGPT) I am willing to complete AI (ChatGPT) related
teaching and training.
training I am willing to teach a class on AI (ChatGPT) to my
intentions students.

45

Authorized licensed use limited to: Zhejiang University. Downloaded on May 01,2024 at 16:06:02 UTC from IEEE Xplore. Restrictions apply.
C. Task Utility Evaluation Tool IV. ANTICIPATED OUTCOMES
While the perception evaluation tool focused on how
teachers would use ChatGPT when teaching students, the The anticipated results of this study are as follows:
evaluation tool based on TAM3 aims to understand how Elementary school teachers' perceptions of utilizing ChatGPT
teachers perceive and think about using ChatGPT in their in education may vary based on their undergraduate major and
professional tasks. Nine variables from the TAM3, aligning the number of hours they have completed in SW·AI training
with the research objectives, were selected. Survey questions over the past two years.
were developed by replacing "System" with "ChatGPT" in the Elementary school teachers teach a variety of subjects, but
existing TAM3 for this purpose. The detailed questions are the during their undergraduate studies, they typically acquire in-
same as those in Table 3. depth knowledge of a particular subject as their major.
Consequently, they tend to utilize their expertise when teaching
TABLE III. TASK UTILITY QUESTIONNAIRE students in their major subject even after graduation. This
Section Question tendency suggests that they might perceive ChatGPT positively
PU1
Using ChatGPT improves my if it is integrated with their specific academic backgrounds and
performance in my job. help them teach their students.
Perceived Using ChatGPT in my job increase my
Usefulness
PU2
productivity.
However, a positive perception of ChatGPT in the classroom
(PU) Using ChatGPT enhances my may not be explained by their academic backgrounds alone. To
PU3
effectiveness in my job. address this aspect, this study included an additional variable,
PU4 I find ChatGPT to be useful in my job. hours of SW·AI training in the past two years. Teachers who
My interaction with ChatGPT is clear and want to utilize SW-AI technology in their classrooms need
PEOU1
understandable.
Perceived Ease Interacting with ChatGPT does not
regular training, as even teachers with the same educational
PEOU2 background can vary in the content of their lessons depending
of Use require a lot of my mental effort.
(PEOU) PEOU3 I find ChatGPT to be easy to use. on which teaching medium they use. Teachers who want to
I find it easy to get the system to do what utilize SW·AI technology in their classrooms are required to
PEOU4
I want it to do.
undergo regular training, as even teachers with the same
PEC1 I have control over using ChatGPT.
Perception of educational background can vary in the content of their lessons
I have the resources necessary to use
External PEC2
Control
ChatGPT. depending on which teaching media they use. Therefore, this
ChatGPT is not compatible with other study aims to reveal the correlation between teachers' specific
(PEC) PEC3
software I use
CANX academic backgrounds, hours of SW·AI training in the past two
ChatGPT do not scare me at all years, and their perceptions of ChatGPT in education. In
1
Computer
CANX
Working with ChatGPT makes me nervous addition, we will also consider supplementary factors such as
2
Anxiety
CANX
gender, teaching experience, and SW·AI educational
(CANX) ChatGPT make me feel uncomfortable. competence to provide additional insights into the anticipated
3
CANX
ChatGPT make me feel uneasy
results.
4

Perceived
ENJ1 I find using ChatGPT to be enjoyable V. CONCLUSION
Enjoyment The actual process of using ChatGPT is
ENJ2
(ENJ) pleasant.
ENJ3 I have fun using ChatGPT. Although there have been studies on how elementary school
VOL1 My use of ChatGPT is voluntary. teachers perceive ChatGPT, these have been approached from a
My supervisor does not require me to use comprehensive perspective rather than analyzing perceptions
VOL2
Voluntariness ChatGPT
based on specific characteristics of elementary school teachers.
(VOL) Although it might be helpful, using
VOL3 ChatGPT is certainly not compulsory in my Currently, there is no research on perceptions of ChatGPT
job. among elementary school teachers according to their individual
REL1 In my job, usage of ChatGPT is important. attributes. Once the results of this study are available in the
Job Relevance REL2 In my job, usage of ChatGPT is relevant. future, we can gain insights into how elementary school teachers
(REL) The use of ChatGPT is pertinent to my
REL3
various job-related tasks.
view ChatGPT based on their profiles. The analysis will serve
I have no difficulty telling others about the as basic factors for utilizing ChatGPT in education, tailored to
RES1
results of using ChatGPT. the characteristics of primary school teachers. Finally, we aim to
I believe I could communicate to others the develop an appropriate technology acceptance model that can
RES2
Result consequences of using ChatGPT
Demonstrability The results of using ChatGPT are apparent
assess the intention to adopt ChatGPT within the framework of
RES3 technology acceptance models developed so far.
(RES) to me.
I would have difficulty explaining why
RES4 using ChatGPT may or may not be
beneficial.
Assuming I had access to the ChatGPT, I REFERENCES
BI1
intend to use it.
Behavioral
Given that I had access to ChatGPT, I
Intention BI2 [1] G. Eysenbach, “The Role of ChatGPT, Generative Language Models, and
predict that I would use it
(BI) I plan to use ChatGPT in the next <n> Artificial Intelligence in Medical Education: A Conversation With
BI3
months

46

Authorized licensed use limited to: Zhejiang University. Downloaded on May 01,2024 at 16:06:02 UTC from IEEE Xplore. Restrictions apply.
 ChatGPT and a Call for Papers,” JMIR Medical Education, vol. 9, no.
1, pp. 1-13, June 2023.
[2] S. Lee, “GPT Generation: how ChatGPT will change the future of our
humanity,” bookmoment, 2023.
[3] I. Kwon, “A Study of the way Forward for Christian in the New Age of
Artificial Intelligence – Focusing on Intergral Epistemology – with
ChatGPT,” Presbyterian Theological Seminary General School, August
2023.
[4] S. Hong, and H. Han, “Analyzing Perception and Educational Needs of
Elementary School Teachers for Using ChatGPT in Education, vol. 26,
no. 4, pp. 51-63, July 2023.
[5] F. Davis, R. Bagozzi, and P. Warshaw, “User acceptance of computer
technology: a comparison of two theoretical models, Management
Science, vol. 35, no.8, 982-1003, August 1989.
[6] V. Venkatesh, and F. Davis, “A theorectical extension of the technology
acceptance model: Four longitudinal field studies, Management Science,
vol. 46, no. 2, 186-204, February 2000.
[7] V. Venkatesh, and H. Bala, “Technology acceptance model 3 and a
research agenda on interventions, Decision Science, vol. 39, no. 2, 273-
315, May 2008.
[8] M. Ryu, and S. Han, “The Educational Perception on Artificial
Intelligence by Elementary School Teachers,” Journal of the Korean
Association of Information Education, vol. 22, no. 3, pp. 317-324, June
2018.
[9] S. Lee, “Elementary School Teachers’ Understanding and Awareness of
AI Education,” The Journal of Korea Elementary Education, vol. 31, pp.
15-31, August 2020.
[10] K. Kim, H. Kwon, and H. Han, “The Analysis of Elementary School
Teachers’ Perception of Using Artificial Intelligence in Education,”
Journal of Digital Convergence, vol. 18, no. 7, pp. 47-56, July 2020.
[11] H. Kim, “Secondary Teachers’ Perception on AI and AI Education,”
Graduate School of Education, Gyeongin National Universtiy of
Educaiton, August 2023.

47

Authorized licensed use limited to: Zhejiang University. Downloaded on May 01,2024 at 16:06:02 UTC from IEEE Xplore. Restrictions apply.