Journals

How ChatGPT is used in higher education, and perceived benefits, risks, and concerns

ChatGPT education technology higher education learning styles student perspectives teacher perspectives

Abstract: The use of generative AI, and in particular ChatGPT, since its launch in November 2022, and other chatbots in the higher education sector has been widespread. This paper explores how students use ChatGPT and other chatbots, their expectations, attitudes and ethical concerns, incorporating findings on attitudes and usage taken from recent literature published since the launch of ChatGPT. The purpose is exploratory, a questionnaire survey of business school students in five countries, USA, Cambodia, Finland, Ghana, and Nigeria, in four continents, during April 2024. The analysis presents and compares opinions from students on their experiences and attitudes towards AI from a learning perspective. The results support the findings that ChatGPT is seen as an assistant providing benefits to personalised learning, whilst recognising the concerns that the overuse of AI presents. In particular, the balance between productive use and ethical concerns is recognised by students. The contribution of this survey is to incorporate the use of and attitudes towards ChatGPT and other chatbots put forward in the recent literature, and to the assessment of the evolution of attitudes to AI in higher education.

Keywords:  ChatGPT, education technology, higher education, learning styles, student perspectives, teacher perspectives

Authors:

Kenneth Paul Charman, CamEd Business School, Phnom Penh, Cambodia, https://orcid.org/0009-0004-3204-3010

Janet Tan, corresponding author, San Jose State University, San Jose, California, USA, janet.tan (at) sjsu.edu, https://orcid.org/0000-0001-5860-3662, 

Sunday Olaleye, Jamk University of Applied Sciences, Finland, https://orcid.org/0000-0002-0266-3989

1. Introduction

Artificial intelligence (AI), a potentially disruptive technology, has increasingly become part of the public and commercial domain as computing technology has evolved. ChatGPT, launched by OpenAI in November 2022, became the first AI chatbot to be available for mass-market personal computers and has led the charge to develop AI applications in virtually every sector, extending the power of computing beyond the evolutionary stage of searching for information to writing reports and assignments, finding and using formulae, and producing finished work which can both save time and extend capacity. Higher education is one sector which has attracted a lot of attention, with AI chatbots perceived as a tool to help with homework for the student, prepare the class for the teacher, and write a paper for the researcher. Tan et al. (2023) first undertook a survey in the spring of 2023 when ChatGPT was at the early adopter stage. This survey was undertaken one year later in the spring of 2024 when students, teachers and researchers had now had greater opportunities to experience ChatGPT and other AI chatbots. This created an opportunity to assess the impact of ChatGPT, for which purposes it was being used, perceived benefits, risks, and ethical concerns. Several studies have been undertaken, highlighting the growing interest in the application of ChatGPT and other AI chatbots, including those studies which have captured the benefits of ChatGPT to students and the potential to change the way that students can learn (Firat 2023; Hasanein & Sobaih 2023; Imran & Almusharraf 2023; Silva et al. 2024), and the benefit to teachers (Rathore 2023). Equally, studies have identified the limitations of ChatGPT, including the potential limits to understanding (Rane et al. 2023), the inability to identify context (Kocoń et al. 2023) and the potential for bias (Adeshola & Adepoju 2023). Researchers have also identified the potential ethical concerns of using ChatGPT (Lund & Wang 2023; Nugroho et al. 2023; Ray 2023). Further research has begun to put forward frameworks for the responsible use of AI (Halaweh 2023; Vázquez-Cano et al. 2023). Research on how students use ChatGPT and their attitudes to using ChatGPT and other AI chatbots has been addressed (Abdaljaleel et al. 2024). But there are still relatively few studies in this area. This study addresses the need for insight into the general applications of ChatGPT and the concerns of students using ChatGPT and contributes to each of the above areas of perceived benefits, risks, ethical considerations, and attitudes towards using ChatGPT and AI chatbots.

The research objective focused on the attitudes of students, teachers, and researchers towards ChatGPT and other AI chatbots. Research questions addressed the uses of ChatGPT and AI chatbots, as well as student perceptions of benefits, risks, and ethical issues related to using AI chatbot technology, which can generate finished or near-finished work for students, teachers, and researchers. The study drew from literature published mainly in 2023 and 2024, which assessed the uses and perceptions of ChatGPT and AI chatbots in the higher educational sector. The survey invited responses from undergraduate and graduate students (and teachers) from five countries (USA, Cambodia, Finland, Ghana and Nigeria) covering four continents (North America, Asia, Europe, and Africa). As an empirical study, it builds on the current research by addressing questions asked by recent researchers in all the above areas of the usage of ChatGPT by students, their attitudes, and ethical concerns. It identifies the extent to which each of the issues raised by recent literature is supported by the survey in the five countries where the survey was undertaken.

The following section provides an overview of the current literature on the benefits of ChatGPT in higher education, the potential limitations and risks of using ChatGPT and other AI chatbots, and the potential changing teaching practices that result from using ChatGPT and AI chatbots. Also included are the current directions and recommendations for research and assessment of the contribution of the current literature to the use and perception of ChatGPT and other AI chatbots. The methodology of the study is then set out and the results are presented and discussed. Conclusions are drawn as to how this study has contributed to the literature, in particular by identifying which of the findings of recent research have been supported by the study, and how this can contribute to the building of a framework to assess the use that students in higher education have made from ChatGPT, their attitudes and concerns towards using ChatGPT and other AI chatbots.

2. Literature review

2.1 Benefits of ChatGPT in higher education

ChatGPT has been cited as applicable to many sectors (Baber et al. 2023; Ray 2023; Zamfiroiu et al. 2023). Several authors outline the scope and usage of AI. These include Zamfiroiu et al. (2023) who, in a systematic literature review, set out the areas of use of AI in educational, medical, writing, and finance and investment settings. Baber et al. (2023) also mentioned the health sector and added legal writing and software coding to the list of uses of AI.

However, higher education stands out (Imran & Almusharraf 2023). The higher education sector lends itself to the use of ChatGPT by the sheer volume of work required which can be applied to ChatGPT. The specific benefits of ChatGPT for students and researchers include personalised learning, greater efficiency in learning, self-directed learning, and the ability to write letters and long and short essays (Firat 2023; Imran & Almusharraf 2023). Teachers could use AI for preparing presentations, grading papers, and building quizzes and exams (Rathore 2023). Several authors have focused on the higher education sector for the extensive use of AI and ChatGPT. Imran and Almusharraf (2023, p. 10) found that “ChatGPT in academia was the most debated topic” with the ability of AI chatbots to write essays, poems, and letters, generate ideas, and provide greater efficiency and learning as an assistant for researchers, teachers, and students. Rathore (2023) extended this to find that teachers could create more engaging presentations and teaching material using AI. Firat (2023) proposed the uses of ChatGPT in self-directed learning, including personalised support, real-time feedback, and added convenience and flexibility in learning. The literature agrees that the scope for ChatGPT and AI chatbots in higher education is broad.

Silva et al. (2024), in an assessment of the use of ChatGPT in the development of software programming in higher education in Brazil, found that human-like responses proved to be a positive benefit, providing a personal assistant-type role, and generating highly accurate responses, allowing for both efficiency and innovation. The ability to create summaries, address queries, and assist with coding assignments was very positive. However, the authors noted the risks of students becoming too dependent on AI-generated code, potentially hindering long-term sustainable educational development. The net result was that most students showed a keen interest in employing ChatGPT and AI chatbots as supplementary resources, providing further support for using ChatGPT and AI chatbots as part of a sustainable and enhanced learning experience.

2.2 Risks and limitations of ChatGPT and AI chatbots in higher education

The perceived benefits are countered by the argument that ChatGPT demonstrated limits to its understanding (Rane et al. 2023), and judging mood or context (Kocoń et al. 2023). ChatGPT and AI chatbots also showed a potential for bias (Adeshola & Adepoju 2023), and were vulnerable to outside threats (Bozkurt & Sharma 2020). Okulu & Muslu (n.d.) experimented with ChatGPT to prepare teaching materials and found both time-saving advantages but limitations to accuracy and the ability to remain focused on the core of the question or learning outcomes. Furthermore, the risks of using ChatGPT and AI chatbots included a lack of accountability for quality and diminished scope for critical thinking (Rane et al. 2023). The literature supported the view that ChatGPT and AI chatbots were best used as assistants (Imran & Almusharraf 2023), with an assessment of the benefits and risks provided by authors such as Silva et al. (2024), and Fütterer et al. (2023).

Rane et al. (2023) focused on the limitations of ChatGPT in playing the role of an author, due to its limited ability for ‘understanding’, and limited scope in critical thinking, experience, adaptability and potential for bias, further hindered by the lack of access to current research and peer review capability. Kocoń et al. (2023) classified ChatGPT as a general AI product but not a master of any one use of AI.

2.3 Changing teaching practices and ethical concerns

Several papers considered ChatGPT to be useful, indeed powerful, but to have ethical concerns. These included Lund & Wang (2023), who proposed that ethical issues such as bias, privacy, autonomy and informed consent, transparency and accountability, as well as intellectual property, were at risk. They even use ChatGPT itself to raise and discuss questions related to its use and ethical issues. Ray (2023) identified many areas of use for ChatGPT and AI chatbots but found that ethical issues were of major concern.

Hasanein and Sobaih (2023) found that ChatGPT has already led to significant changes in methods of learning and teaching, providing more conversational learning and more human-like responses. The authors utilised Constructivism Learning Theory (Nurhuda et al. 2023) to explore the motivations underlying higher education students in their use of ChatGPT, and the consequences for teaching and learning. They expressed concern for the ethical implications of using AI and suggested clear policies, guidelines and training aimed at responsible use.

Alenezi et al. (2023) demonstrated how ChatGPT was significantly changing how English as a foreign language was taught. Nugroho et al. (2023) supported the potential of ChatGPT for language learning but also pointed out the ethical concerns. Halaweh (2023) led the direction of building strategies for responsible implementation of AI in the higher education sector. The literature recognised the changes in teaching practices brought about by incorporating the use of ChatGPT and AI chatbots, and that these changes in teaching and learning practices came with practical and ethical concerns.

2.4 The attitudes of students and educators to ChatGPT

Given the substantiation benefits but also risks of relying more on AI chatbots such as ChatGPT, researchers have begun to examine the attitudes of students and educators toward adopting AI technology. Vázquez-Cano et al. (2023) were one of the first to provide an empirical assessment on the impact of AI, and again this included both methodology and ethical issues. The authors proposed that any debate on the integration of AI must begin with assessing how AI can best be used to provide a productive yet responsible way to support learning and educational and professional development for students and the wider population.

Abdaljaleel et al. (2024), undertook a multinational study on the factors influencing university students’ attitudes to ChatGPT in Arab countries. They found that nearly half of the respondents were aware of ChatGPT and had used it before the survey. Ease of use, social influences, behavioural and cognitive influences, low perceived risks and low anxiety were common among students surveyed. However, there were differences by country, age, university type, and academic performance. The study again recommended context-specific guidance on ChatGPT usage, taking into account the different attitudes of students towards ChatGPT across countries.

Cooper (2023) explored the use of ChatGPT as generative AI in science education and found that generative AI offers transformational potential in science education, providing for designing science units, rubrics, quizzes, and the like. However, the study noted that educators should ensure and model responsible use of ChatGPT including prioritising critical thinking, clearly communicating expectations, and adapting the use of ChatGPT and AI in teaching contexts. The role of ChatGPT and AI in science education was foreseen but with an emphasis on responsible use.

Fütterer et al. (2023) assessed broadly the impact of ChatGPT in higher education, and this can sum up many of the findings from the recent literature. Attitudes varied from enthusiastic to identifying concerns about restricting learning opportunities and potentially contributing to misinformation. The authors analysed Twitter (now X) data to find that the education sector was the source of a high proportion of sharing of perceptions and reactions to ChatGPT. Issues such as cheating were raised and mixed reactions to the use of ChatGPT, both positive and negative were common.

Further to the cognitive issues of how ChatGPT was changing (and improving) learning, there was the ethical question of whether ChatGPT was breaking ethical boundaries through the potential for bias, discrimination, endangering academic integrity, and facilitating plagiarism (Lund & Wang 2023; Ray 2023). Several authors advocated for the introduction of clear policy, guidelines, and training for the use of ChatGPT in all areas of education (Abdaljaleel et al. 2024; Fütterer et al. 2023; Hasanein & Sobaih 2023; Silva et al. 2024; Vázquez-Cano et al. 2023).

2.5 Potential applications and future directions

Researchers have put forward assessments of trends and directions in the use of ChatGPT and similar AI applications in education. Rahman & Watanobe (2023) provided a comprehensive assessment of the opportunities for using ChatGPT, combined with the threats, and identified strategies being used to incorporate ChatGPT in education from the viewpoint of educators and students. They identified such issues as personalised feedback, improvements in accessibility, interactive learning, lesson preparation, assistance in evaluation, and new teaching methods as opportunities, but highlighted the potential for cheating and reduced critical thinking as challenges. This was augmented by the lack of ability to evaluate or assess information generated by ChatGPT. 

Iskender (2023) exemplified a growing trend in authors to use ChatGPT as an ‘interviewee’ to explore its impact on higher education and in particular academic publishing, with the AI response claiming that it could allow instructors to spend more time on more intellectually demanding activities, and at the same time provide students with brainstorming opportunities. Keiper et al. (2023) explored the potential for ChatGPT in sports management education, finding that ChatGPT responded to qualitative inquiries with short answers which were grammatically correct, complete, and accurate, and offered suggestions as to how AI chatbots such as ChatGPT could augment sports management education. Meron & Araci (2023) examined the feasibility of using ChatGPT as a virtual assistant in designing materials for higher education design students and creating course materials. They found that using ChatGPT saved time, and provided structured content, thus enabling brainstorming. However, the content was found to be generic, requiring subsequent extensive editing and augmentation by a human hand. In short, the authors found potential for use in course design, but with functional and ethical limitations.

Farrokhnia et al. (2024) undertook a SWOT analysis of ChatGPT and its implications for educational practice and research. They found strengths to include a sophisticated language model with capabilities for self-improvement, and personalised real-time responses, providing an effective role as an assistant. This provided users with opportunities to enhance access to information, facilitating personalised learning, a developing theme of several authors. They also found that ChatGPT reduced teaching workload, a major constraint on teaching. However, they identified weaknesses in that ChatGPT lacked a deep understanding, risked bias, led to difficulties in evaluating quality, and reliance on ChatGPT tended to limit higher-order thinking skills. These combined to provide threats to education including a lack of contextual understanding, facilitating plagiarism, diminishing high-order skills, and had the potential to perpetuate discrimination and endanger academic integrity. This provided further support for research to help develop guidelines for educational practices for using ChatGPT and other AI chatbots.

2.6 Conclusion of the literature review: Identifying the gaps

Surveys of student perceptions and attitudes toward ChatGPT and AI chatbots remain relatively few, and this is where we consider that there are research gaps. The literature has identified many benefits of using ChatGPT and AI chatbots, and overwhelmingly these place ChatGPT in the role of assistant and co-pilot. However, the challenges to traditional education systems, such as potential cheating, human-like text generation, reduced critical thinking skills and difficulties in evaluating the quality of ChatGPT-generated education have raised fundamental limitations and ethical concerns about using AI in education. These benefits, limitations, risks, and ethical concerns are perceived by both educators and students. Our paper contributes to the growing body of literature on attitudes, particularly of students using ChatGPT and AI chatbots, and more widely AI, and how attitudes are evolving.

3. Methodology

3.1 User surveys

The research objective was to explore the benefits students (and educators) perceive and value using ChatGPT and AI chatbots, and the limitations and risks that using ChatGPT and AI chatbots presents. This research was based on survey analyses with Likert scale questionnaires. The survey was undertaken during the spring of 2024, 18 months after the launch of ChatGPT. The research questions were exploratory. Firstly, how students were using ChatGPT, and how they intended to use ChatGPT in the future (See Appendix question 5). Secondly, the impact of learning outcomes. Thirdly, students’ perceived benefits of using ChatGPT and AI chatbots (See Appendix Question 8). Fourthly, what students saw as the downsides of using ChatGPT (See Appendix Question 9). The authors used the opportunity to see whether there were any significant differences between the responses of students in different countries. However, as this research was fundamentally exploratory, the authors did not set out specifically to identify differences between countries. The regional differences which were revealed by the study are discussed in the results section.

3.2 Data collection

The study employed questionnaires to business school students and faculty members in North America, Europe, Asia, and Africa. The basis of the survey was non-probability convenience sampling, using student populations which were available to the researchers, which would likely provide a high response rate, and would enable the survey to be undertaken and completed in a relatively short time frame. Convenience sampling is characterised by the closeness of the target audience to the researcher and recognises that data cannot be generalised beyond the source of the group, consistent with the assessment of convenience versus purposeful sampling (Obilor 2023). However, focusing on undergraduate and graduate business school students does provide a population which has reason to use ChatGPT and AI chatbots, can provide valuable insight into the use and perceptions of ChatGPT and AI chatbots from this population, and builds on the 2023 survey which addressed the same target population. As the nature of this research was fundamentally exploratory it was decided that convenience sampling would be appropriate as this significantly reduces both the cost of the research and the lapsed time of the survey, a choice consistent with attributes of non-probability sampling (Bell et al. 2022, pp. 199–201).

3.3 Data analysis

The approach with both Likert scale and open-ended questions was deemed appropriate due to the rapidly evolving nature of generative AI usage, and is also consistent with the approach used in other surveys (Abdaljaleel et al. 2024; Hasanein & Sobaih 2023; Rahman & Watanobe 2023; Silva et al. 2024). The study was cross-sectional, undertaken in April 2024. The survey was based on the issues raised by the recent literature and structured in line with the 2023 survey. The data was gathered from April to May 2024. Most questions related to the usage of and perceptions of ChatGPT were Likert scale questions. According to Johns (2010), Likert scale questions are suitable to measure attitudes and values. The questionnaire (See Appendix) was piloted in March 2024 and checked for reliability through the reliability test of Cronbach Alpha (Peterson 1994). Cronbach’s alpha is a measure used to assess the reliability or internal consistency of a set of survey or test items. It is a tool that helps researchers check if a set of questions or items on a test are all working together to measure the same thing. Based on the 0.7 level threshold of Cronbach Alpha, the questionnaires were adjusted to suit the study objectives. The final data from Google Forms was cleaned to meet the assumption for the target data analysis.

4. Survey results and interpretation

4.1 Location and study focus

The survey comprised students from universities in five locations covering four continents. The total sample size was 466, comprising 45.5% from Asia (mostly Cambodia), 30.7% from North America (mostly California), 15.7% from Africa (mostly Ghana and Nigeria), and 8% from Europe (mostly Finland), see Figure 1.

Of the total of 466 respondents, 78.1% were undergraduate students and 15.2% graduate students. A total of 4.0% were instructors and tutors (including students who taught part-time) and 2.0% were technical support staff or administrators. Only one student (0.25%) was a researcher, see Figure 2.

For the surveyor’s study focus, there are 61.8% Business & Management, 21.2% Accounting & Finance, 4.5% Science & Mathematics, 3.4% Humanity & Social Science, 3.4% Engineering, and 3.0% Medical Field, and 2.7% others, see Figure 3.

4.2 Awareness of ChatGPT and AI chatbots

Questions required responses on a scale ranging from one to five, equating to “1. not at all”, “2. a little”, “3. a reasonable amount, enough (or sufficient)”, “4. quite a lot”, or “5. very well / very much”. A total of 17.4% of respondents stated that they knew ChatGPT/AI chatbots very well, 29.8% well, 35.8% moderately or sufficiently well. Only 10.7% of respondents stated that they knew about ChatGPT/chatbots a little, and 6.2% not at all, see Figure 4. A higher proportion of respondents in Europe (71.1%) and Africa (54.8%) claimed to know ChatGPT/chatbots quite a lot & very well compared with (43.4%) in both North America and Asia (42.9%). In Africa the highest proportion of respondents (16.4%) claimed not to know about ChatGPT at all. In Asia 10% of respondents knew ChatGPT/chatbots not at all, North America 7%, and no-one in Europe, see Figure 5.

4.3 Current purposes of using ChatGPT/chatbots

The highest number of respondents those who stated that they used ChatGPT regularly, often or very often used ChatGPT/chatbots for general knowledge (70.8%), Search information and ideas (70.4%), brainstorming (69.7%), research (69.3%), answering questions (68.2%), homework help (65.9%), studying (65.0%), and giving a second opinion (62.4%). A lower number of respondents used ChatGPT for obtaining formulae (e.g. for statistics) (45.9%), searching for Literature Resources (45.9%), writing essays (44.2%), and writing reports (43%), entertainment (38.8%), and teaching (38.4%), see Figure 6 for the percentage breakdown.

Respondents referred to specific uses, including search and general knowledge uses, such as encouraging debate, finding and developing ideas, exploration, general knowledge, Planning, presentations, search engine, summarising text, providing insight, help with studying, modification of language, and paraphrasing. More technical uses included the analysis of data, cross-checking information, feedback and proofreading, and translating documents. Personal administration and personal growth use included: art, cooking, creative activities, and entertainment. Some specific comments from respondents included using ChatGPT for:

  • “Brainstorming, good ideas, researching information that I don’t understand”
  • “I can ask it in all kinds of ways”
  • “Random question when I feel like chatting with a person whom I do not know”
  • “Roadmap for any task with structure”
  • “Especially using to brainstorm, good ideas, researching information that I don’t understand”
  • “To summarise a context which is too long to read”
  • “Correcting grammar and writing”
  • “Drafting a business plan, mission, and value statements”

4.4 Potential benefits of using ChatGPT/AI chatbots for learning purposes

The majority of the respondents (56.7%) considered that ChatGPT helped them achieve a better or much better learning outcome, see Figure 7. A total of 44.7% of respondents had been given moderate or more guidance on ChatGPT, so it has been more a case (40.6%) of finding their own way, see Figure 8.

The most widely considered high and very high benefits were in terms of access to information (59.7%), improved understanding of complex concepts (51.5%), helping generate ideas and brainstorming (51.3%), and facilitating real-time feedback for current work (41.8%).  Respondents saw ChatGPT as a personal assistant (41.8%), as an aid to help write letters, reports, and essays (36.9%), to facilitate work evaluation (35.6%), to provide a personalised way of learning which suits me (34.1%). Increased motivation for learning was also cited as a benefit (31.1%), as was improved critical thinking (28.8%), and providing more human-like responses or interaction (26.8%), see Figure 9.

Respondents referred to specific benefits in their open-ended comments, including ease of access to information, efficiency, generating ideas, and alternative answers, summarising text, concluding concepts, ability to translate quickly, and the ability to check for grammar and mistakes, paraphrasing, and giving a second opinion. Some poignant comments on the benefits of ChatGPT and AI chatbots:

  • “Can help in brainstorming”
  • “For mathematics, it explains in detail how to solve problems which is really helpful when you have been stuck on it and you don’t know what formula to use”
  • “Help how to use and explain all the questions in an essay.”
  • “Improve my work more quickly and effectively.”
  • “It spurs me to read more”
  • “It’s faster and reduces waste of time”
  • “Motivation and increase the passion when we not sure in decision making”
  • “Sometimes Google just doesn’t provide that answer. Even if ChatGPT may not always have the best answers, it is the best available resource.”

4.5 Potential drawbacks (and ethical issues) of ChatGPT/chatbots

The most widely cited potential drawbacks (considered as high and very high risks) were: the risk of plagiarism (60.0%), dependence on technology (51.5%), reductions in critical thinking (50.6%), lack of peer review (48.7%), lack of personal interaction with instructors and peers (44.8%), possible biased responses (43.3%), inability to assess accurately (37.3%), and lack of access to up-to-date research (35.4%), see Figure 10.

Open-ended comments on the potential drawbacks of ChatGPT included the potential for: data privacy and security issues, risks of plagiarism, dependence on technology, and inaccurate answers. Respondents also referred to the lack of up-to-date information, lack of control, lack of credibility, lack of in-depth knowledge and originality. Lack of personal interaction with peers and professors and reduction in problem-solving ability were also considered downsides by respondents. Comments from respondents on possible downsides and ethical issues of using ChatGPT and AI chatbots included:

  • ‘The results are often plagiarism-free, therefore it’s hard to find the source of information.’
  • “It gives a different answer when we ask it again”
  • “ChatGPT could possibly just give a complete nonsense answer”
  • “It might produce inaccurate information and our critical thinking will be lower”
  • “Reduces the discovery process, which would hinder learning”
  • “Reduction in problem-solving ability”
  • “The inaccuracy in the citation is a potential drawback”
  • “We do not know the date of the data they gave us”
  • “The disadvantage of using ChatGPT is diluting the human creativity”

5. Summary discussion of survey results

The descriptive survey results provide a comprehensive overview of the demographic characteristics, awareness, usage patterns, benefits, and potential drawbacks of ChatGPT among students and educational staff. By analysing these data points, several themes and insights emerge that can inform the effective integration of ChatGPT into educational settings.

5.1 Awareness of ChatGPT/chatbots

Awareness of ChatGPT was generally high among the respondents, with 17.4% of students claiming to know ChatGPT very well and 29.8% well. A significant proportion (35.8%) had a moderate understanding. However, 10.7% needed more knowledge of ChatGPT, and 6.2% stated they had little or no awareness of ChatGPT. Regional variations in awareness were observed, with higher awareness in Europe (34.2% very well) and Africa (54.8% very well), but also a significant portion of students in Africa (16.4%) were unaware of ChatGPT.

5.2 Purposes of use for ChatGPT/AI chatbots

Students used ChatGPT for various educational purposes, reflecting recent literature findings. The most common themes for moderate, high, or very high uses included general knowledge (70.8%), followed by research information (70.4), brainstorming (69.7%), research (69.3%), answering questions (68.2%), homework help (65.9%), studying (65.0%), and giving a second opinion (62.4). These uses indicate that students primarily utilised ChatGPT as an assistant for general knowledge acquisition and idea generation rather than specific task completion. Less frequent but still substantial uses included obtaining formulas (45.9%), searching for literature resources (45.9%), writing essays (44.2%), and writing reports (42.9%). These results suggest that while many students leveraged ChatGPT to generate content and research support, a smaller proportion relied on it to produce finished academic work. Entertainment (38.8%) was ranked last.

5.3 Benefits of ChatGPT/AI chatbots

The survey revealed several key benefits considered moderate, high, or very high of using ChatGPT in higher education. Quick access to information (88.4%) topped the list, followed by generating ideas and brainstorming (83.0%), improved understanding of complex concepts (82.2%), real-time feedback for current work (74.5%), a personal assistant role (71.2%), facilitates work evaluation (70.4%), personalised learning (70.2%), increased motivation for learning (68.0%), help to write letters, reports, and essays (68.0%), improved critical thinking skills (60.7%), and more human-like responses and interaction (60.1%). These benefits underscored ChatGPT’s role in enhancing access to information and supporting comprehension and creativity. Students perceived it as a valuable tool for facilitating work evaluation and providing personalised learning experiences, aligning with its use as an educational assistant.

Most students considered that ChatGPT provided them with a better learning outcome. So even though they may not have used it a large amount, and even though using it only as an assistant, often as a secondary assistant, students still felt that ChatGPT or other chatbots were useful for their learning.

5.4 Potential drawbacks of ChatGPT/AI chatbots

Despite the benefits, the survey highlighted several potential drawbacks considered to be moderate, high, or very high risks. The risk of plagiarism (85.6.0%) was the most notable drawback, followed by dependence on technology (81.3%), reduction in critical thinking (79.4%), lack of peer review (76.6%), possible biased responses (75.5%), inability to assess accurately (73.4%), lack of personal interaction with instructors and peers (71.9%), and lack of access to up-to-date research (63.9%). These concerns point to significant challenges in the educational use of ChatGPT. The risk of plagiarism and reduced critical thinking skills are particularly worrying, as they can undermine academic integrity and students’ intellectual development. Additionally, the reliance on technology and potential response biases highlights the need for critical evaluation and proper guidance when using AI tools.

5.5 Themes and insights

The survey results reveal several themes: educational enhancement, support role, awareness and accessibility, ethical and cognitive concerns, and integration and guidance. ChatGPT has been widely recognised for its ability to enhance learning by providing quick access to information, improving understanding of complex topics, and aiding in brainstorming and idea generation. The tool was primarily used as an educational assistant rather than for completing specific academic tasks, indicating its supportive role in learning processes. Regional disparities existed in awareness and use of ChatGPT, suggesting a need for more comprehensive training and accessibility to ensure equitable benefits. Potential drawbacks, such as plagiarism, dependency on technology, and reduced critical thinking skills, emphasise the importance of developing ethical guidelines and promoting the balanced use of AI in higher education. Evidently, the high perceived benefits combined with significant risks point to the necessity of structured integration of ChatGPT in educational curricula, accompanied by adequate guidance and support for students and higher education educators.

Given that the results are consistent with the current literature we suggest more rigorous studies can now be undertaken to confirm or otherwise the findings of attitudes that have been found.

6. Conclusion

The survey results indicate that while ChatGPT offers substantial benefits for enhancing learning and supporting educational activities, critical concerns must be addressed. These include ethical considerations, ensuring the development of critical thinking skills, and providing sufficient guidance to mitigate the risks associated with AI use in higher education. Future research and policymaking should focus on creating frameworks that maximise ChatGPT’s educational advantages while addressing its potential drawbacks.

This study provides insights into the use and perceptions of ChatGPT among higher education students in various countries, focusing on its perceived benefits, risks, and ethical concerns. The findings have several practical implications for educators in higher education. First, educators need to integrate AI tools like ChatGPT to enhance personalised learning while mitigating the risk of over-reliance. Institutions could develop workshops or modules to guide students on using ChatGPT effectively for brainstorming, research, and idea generation, but without compromising critical thinking and academic integrity. Additionally, incorporating AI ethics into curricula is essential, as many students highlighted the ethical risks associated with ChatGPT, including plagiarism and bias. Comparing our findings with earlier studies, we observe some similarities and notable differences. Consistent with prior research, such as Imran and Almusharraf (2023), students in this study largely view ChatGPT as a valuable educational assistant, enhancing their efficiency in completing tasks and understanding complex concepts despite its challenges. However, unlike earlier studies that focused on the technical field (Silva et al. 2024), our study captures a broader range of literature, where the ethical concerns surrounding AI usage are underscored. Furthermore, our cross-continental survey reveals regional differences in awareness and application of ChatGPT, which were less emphasised in earlier research.

This study, however, has certain limitations. The use of non-probability convenience sampling limits the ability to generalise the results beyond the sampled population. Additionally, the survey’s focus on business school students, though relevant, may only partially capture the broader educational landscape. Furthermore, as an exploratory study, the research primarily identifies trends without delving deeply into specific causal relationships or longitudinal impacts of ChatGPT use.

Future research should address these limitations by employing more robust sampling methods and exploring different educational settings beyond business schools. Longitudinal studies provide a deeper understanding of how students’ attitudes and usage of AI tools like ChatGPT evolve. Additionally, research into the long-term effects of AI on critical thinking skills, academic integrity, and learning outcomes would be valuable. Expanding research into regional differences in AI adoption can also provide insights into how socio-cultural factors influence the integration of AI in higher education.

References

Abdaljaleel, M., Barakat, M., Alsanafi, M., Salim, N. A., Abazid, H., Malaeb, D., Mohammed, A. H., Hassan, B. A. R., Wayyes, A. M., Farhan, S. S., Khatib, S. E., Rahal, M., Sahban, A., Abdelaziz, D. H., Mansour, N. O., AlZayer, R., Khalil, R., Fekih-Romdhane, F., Hallit, R., … Sallam, M. (2024). A multinational study on the factors influencing university students’ attitudes and usage of ChatGPT. Scientific Reports, 14(1), 1983–1983. https://doi.org/10.1038/s41598-024-52549-8

Adeshola, I., & Adepoju, A. P. (2023). The opportunities and challenges of ChatGPT in education. Interactive Learning Environments, 0(0), 1–14. https://doi.org/10.1080/10494820.2023.2253858

Alenezi, M. A. K., Mohamed, A. M., & Shaaban, T. S. (2023). Revolutionizing EFL special education: how ChatGPT is transforming the way teachers approach language learning. Innoeduca, 9(2), 5–23. https://doi.org/10.24310/innoeduca.2023.v9i2.16774

Baber, H., Nair, K., Gupta, R., & Gurjar, K. (2023). The beginning of ChatGPT – a systematic and bibliometric review of the literature. Information and Learning Science, 125(7/8), 587-614. https://doi.org/10.1108/ILS-04-2023-0035

Bell, E., Bryman, A., & Harley, B. (2022). Business research methods. Oxford university press.

Bozkurt, A., & Sharma, R. (2020). Emergency remote teaching in a time of global crisis due to CoronaVirus pandemic. Asian Journal of Distance Education, 15, i–vi. https://doi.org/10.5281/zenodo.3778083

Cooper, G. (2023). Examining Science Education in ChatGPT: An Exploratory Study of Generative Artificial Intelligence. Journal of Science Education and Technology, 32(3), 444–452. https://doi.org/10.1007/s10956-023-10039-y

Farrokhnia, M., Banihashem, S. K., Noroozi, O., & Wals, A. (2024). A SWOT analysis of ChatGPT: Implications for educational practice and research. Innovations in Education and Teaching International, 61(3), 460–474. https://doi.org/10.1080/14703297.2023.2195846

Firat, M. (2023, January). How ChatGPT can transform autodidactic experiences and open education? https://doi.org/10.31219/osf.io/9ge8m

Fütterer, T., Fischer, C., Alekseeva, A., Chen, X., Tate, T., Warschauer, M., & Gerjets, P. (2023). ChatGPT in education: Global reactions to AI innovations. Scientific Reports, 13(1), 15310–15314. https://doi.org/10.1038/s41598-023-42227-6

Halaweh, M. (2023). ChatGPT in education: Strategies for responsible implementation. Contemporary Educational Technology, 15(2), ep421. https://doi.org/10.30935/cedtech/13036

Hasanein, A. M., & Sobaih, A. E. E. (2023). Drivers and consequences of ChatGPT use in higher education: Key stakeholder perspectives. European Journal of Investigation in Health, Psychology and Education (EJIHPE), 13(11), 2599–2614. https://doi.org/10.3390/ejihpe13110181

Imran, M., & Almusharraf, N. (2023). Analyzing the role of ChatGPT as a writing assistant at higher education level: A systematic review of the literature. Contemporary Educational Technology, 15(4), ep464. https://doi.org/10.30935/cedtech/13605

Iskender, A. (2023). Holy or unholy? Interview with Open AI’s ChatGPT. European Journal of Tourism Research, 34, 3414. https://doi.org/10.54055/ejtr.v34i.3169

Johns, R. (2010). Survey question bank: Methods Fact Sheet 1, Likert items and scales. University of Strathclyde, 1, 1–11.

Keiper, M. C., Fried, G., Lupinek, J., & Nordstrom, H. (2023). Artificial intelligence in sport management education: Playing the AI game with ChatGPT. Journal of Hospitality, Leisure, Sport and Tourism Education, 33, 100456. https://doi.org/10.1016/j.jhlste.2023.100456

Kocoń, J., Cichecki, I., Kaszyca, O., Kochanek, M., Szydło, D., Baran, J., Bielaniewicz, J., Gruza, M., Janz, A., Kanclerz, K., Kocoń, A., Koptyra, B., Mieleszczenko-Kowszewicz, W., Miłkowski, P., Oleksy, M., Piasecki, M., Radliński, Ł., Wojtasik, K., Woźniak, S., & Kazienko, P. (2023). ChatGPT: Jack of all trades, master of none. Information Fusion, 99, 101861. https://doi.org/10.1016/j.inffus.2023.101861

Lund, B. D., & Wang, T. (2023). Chatting about ChatGPT: How may AI and GPT impact academia and libraries? Library Hi Tech News, 40(3), 26–29. https://doi.org/10.1108/LHTN-01-2023-0009

Meron, Y., & Araci, Y. T. (2023). Artificial intelligence in design education: evaluating ChatGPT as a virtual colleague for post-graduate course development. Design Science, 9(e30), 24. https://doi.org/10.1017/dsj.2023.28

Nugroho, A., Putro, N. H. P. S., & Syamsi, K. (2023). The potentials of ChatGPT for language learning: Unpacking its benefits and limitations. Register Journal, 16(2), 224–247. https://doi.org/10.18326/register.v16i2.224-247

Nurhuda, A., Al Khoiron, M. F., Syafii Azami, Y., & Nimah, S. J. (2023). Constructivism learning theory in education: Characteristics, steps and learning models. Research in Education and Rehabilitation, 6(2), 234–242. https://doi.org/10.51558/2744-1555.2023.6.2.234

Obilor, E. I. (2023). Convenience and purposive sampling techniques: Are they the same? International Journal of Innovative Social & Science Education Research, 11(1), 1–7.

Okulu, H. Z., & Muslu, N. (n.d.). Designing a course for pre-service science teachers using ChatGPT: what ChatGPT brings to the table. Interactive Learning Environments, 1–18. https://doi.org/10.1080/10494820.2024.2322462

Peterson, R. A. (1994). A meta-analysis of Cronbach’s coefficient alpha. Journal of Consumer Research, 21(2), 381–391. https://doi.org/10.1086/209405

Rahman, M. M., & Watanobe, Y. (2023). ChatGPT for education and research: Opportunities, threats, and strategies. Applied Sciences, 13(9), 5783. https://doi.org/10.3390/app13095783

Rane, N., Choudhary, S., Tawde, A., & Rane, J. (2023). ChatGPT is not capable of serving as an author: ethical concerns and challenges of large language models in education. International Research Journal of Modernization in Engineering Technology and Science, 5, 851–874. https://doi.org/10.56726/IRJMETS45212

Rathore, D. B. (2023). Future of AI & Generation Alpha: ChatGPT beyond boundaries. Eduzone: International Peer Reviewed/Refereed Multidisciplinary Journal, 12(1), Article 1. https://www.eduzonejournal.com/index.php/eiprmj/article/view/254

Ray, P. P. (2023). ChatGPT: A comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope. Internet of Things and Cyber-Physical Systems, 3, 121–154. https://doi.org/10.1016/j.iotcps.2023.04.003

Silva, C. A. G. da, Ramos, F. N., de Moraes, R. V., & Santos, E. L. dos. (2024). ChatGPT: Challenges and benefits in software programming for higher education. Sustainability, 16(3), 1245. https://doi.org/10.3390/su16031245

Tan, J., Charman, K., Domenghino, M., & Olaleye, S. (2023). ChatGPT: A disruptive blessing in disguise for higher education? Finnish Business Review. https://verkkolehdet.jamk.fi/finnish-business-review/2023/08/07/chatgpt-a-disruptive-blessing-in-disguise-for-higher-education/

Vázquez-Cano, E., Ramírez-Hurtado, J. M., Sáez-López, J. M., & López-Meneses, E. (2023). ChatGPT: The brightest student in the class. Thinking Skills and Creativity, 49, 101380. https://doi.org/10.1016/j.tsc.2023.101380

Zamfiroiu, A., Vasile, D., & Savu, D. (2023). ChatGPT–A systematic review of published research papers. Informatica Economica, 27(1), 5–16.

Appendix

This research addresses the following questions.

The first question is whether ChatGPT and more widely generative AI provide a benefit, and the overwhelming answer is that it can save time dramatically, albeit with the need to double-check the quality of the answers provided.   

The next question of howChatGPT is important is that of how the use of ChatGPT changes how students and teachers learn, compared to how they learned before. The personalised learning provided by ChatGPT, the ability to use ChatGPT, or more widely generative AI, to ask and find answers to questions in your way, is an important question (Firat 2023).

Below are the survey questions sent to various higher educational institutions randomly.

Survey Questions:

1. Which region are you in? (North America, Europe, Asia, Latin America, Africa, Other)

2. Who are you? (undergraduate student, graduate student, professor/instructor, technical/support staff/administrator, other)

3. What is your study focus? (science and mathematics, engineering, business and management, humanity and social science, medical field, other)

4. How well do you know about ChatGPT? (Likert scale 1 – not at all to 5 – very well)

5. If you use chatbot, for what purpose do you typically use it? (Check all that apply with Likert Scale rating, 1 – not at all to 5 – very much) (for research, studying, homework help, general knowledge, teaching, brainstorming/idea generation, obtaining formulas – e.g. for statistics, search for literature resources, search information and ideas, answering questions, giving a second opinion, writing essays, writing reports, entertainment, other – please specify below)

6. To what degree do you think ChatGPT/chatbot helps you achieve a better learning outcome? (Likert Scale 1 – not helpful at all to 5- very helpful)

7. Have you received any guidance or training on how to effectively use ChatGPT/chatbot for learning purposes? (Likert scale from 1 – not at all to 5 – very high)

8. In your opinion, what are the potential benefits of using ChatGPT/chatbot for learning? (check all that apply with Likert Scale from 1 – very low to 5 – very high) (quick access to information, improved understanding of complex concepts, increased motivation for learning, improved critical thinking skills, a personalised way of learning which suits me, real-time feedback for current work, facilitates work evaluation, to be a personal assistant, more human-like responses and interaction, to help write letters, reports, and essays, to help generate ideas and brainstorming, other – please specify below)

9. In your opinion, what are the potential drawbacks of using ChatGPT for learning purposes? (Check all that apply with Likert Scale from 1 – very low to 5 – very high)(dependence on technology, lack of personal interaction with instructors and peers, reductions in critical thinking, lack of access to up-to-date research, inability to assess, the risk of plagiarism, lack of peer review, possible biased responses, other – please specify below)

URN: http://urn.fi/urn:nbn:fi:jamk-issn-2341-9938-84

Back to top Back to top