Generative AI: is it a paradigm shift for higher education?

ABSTRACT

In this special issue, we explore the opportunities and challenges of using Generative AI (GenAI), in particular, text generators in higher education learning and teaching. As GenAI has become increasingly popular with many staff and students, this special issue provides an overview of the current state of the field and offers insights into future research. This introduction paper consists of four parts. It begins by providing an overview of AI and Generative AI, identifying the gap and framing the special issue relating to the gaps. The second part explores the opportunities and challenges of GenAI in higher education, as identified in the literature. The third part provides an overview of the papers included in the special issue. The final part is the self-reflection of the lead author. The special issue aims to serve as a valuable resource for higher education stakeholders, such as students, practitioners, researchers and managers. We hope this collection will help advance knowledge and future research, encourage innovation and inform evidence-based policy and practices in the field of Generative AI in higher education.

KEYWORDS:

• Generative AI
• higher education
• AIEd
• ethics
• learning and teaching

Introduction

This special issue for the Studies in Higher Education (SHE), titled ‘AI text generators: a threat, an opportunity or an asset to higher education’ addresses a popular and timely topic – Generative Artificial Intelligence (GenAI) in higher education. GenAI may be considered a subset of Artificial Intelligence in higher education (AIEd). AI is a core technology in digital transformation and is increasingly playing an important role in transforming the way that education is delivered and experienced. AI adoption in learning and teaching in higher education has become a popular research topic in recent decades. The disruptive role of AI, including the challenges and weakness of AIEd, and its applications across various learning and teaching contexts, such as providing students with automated learning and administrative support; predicting student academic performance; and marking student assessments automatedly, have been discussed and evaluated in various systematic review papers (Bearman, Ryan, and Ajjawi 2023; Crompton and Burke 2023; Zawacki-Richter et al. 2019). Nevertheless, the previous research on AIEd has concentrated on either the pre-GenAI era, or using AI to perform a specific task, as outlined above.

The advent of GenAI, such as ChatGPT, has further accelerated AI adoption in higher education (Bozkurt et al. 2023). GenAI software, including text generators, are built on large language models (LLMs), such as Generative Pre-trained Transformer (GPT). LLMs are specific models within natural language processing (NLP) and deal with language-related tasks (O'Dea and O'Dea 2023a). GenAI software can produce text, audio or image-based responses to questions, including relatively complex queries, such as summarising a research paper, language translation and producing creative content (e.g. emails, blog posts, news items, and scientific articles). These responses are very close to and sometimes indistinguishable from human writing and language.

The convenience of GenAI, such as ease of use, large volume of information and the perceived difficulty in accurate detection (Guo et al. 2023), has promoted its usage among university students and subsequently encouraged universities to explore ways of embedding GenAI in learning and teaching. As with other disruptive technologies, such as Big Data and 3D printing, there are conflicting views towards GenAI (Farhi et al. 2023). Some believe that this new disruptive technology can be a valuable asset to enhance learning, particularly in supporting students to improve their academic writing and critical thinking skills. There are also concerns raised with regard to its ethical implications. In fact, some consider GenAI a major threat to academic integrity and believe that students should be banned from using them (Eke 2023).

The rapid growth of GenAI has generated considerable interest in the field of AIEd since November 2022, and there have been peer-reviewed journal articles discussing opportunities and concerns in higher education (Atlas 2023; Michel-Villarreal et al. 2023; Rudolph, Tan, and Tan 2023). However, more empirical studies are needed to explore the adoption of GenAI (e.g. other than ChatGPT) in different country contexts. It is also important and necessary to examine institutional policies for the effective integration of GenAI in teaching and research so that academic and support staff can effectively guide and support the ethical and equitable use of GenAI among students. This special issue aims to help bridge the gaps. It calls for international research on both theory and practices in areas such as learning and teaching activities, academic integrity, ethics, social well-being and critical academic writing. Before exploring the details of the individual papers, we first examine the opportunities and challenges of GenAI in higher education, presented in the current literature.

Opportunities

The rise of GenAI, as discussed below, has raised concerns about academic and research ethics. However, it could also be seen as a unique opportunity to explore ways of using such disruptive technology for moving teaching and learning practices forward. This is because AI technologies, such as facial recognition, AI-powered search engines and virtual agents, are already an integral part of our work and lives and will continue reshaping the future of businesses (Soni et al. 2019). By experimenting with the tool, Kohnke and colleagues (2023) offered recommendations on how ChatGPT may be used in language teaching and learning, including creating simulated speaking environments, and generating self-test quizzes for students. In the context of computer science education, MacNeil and colleagues (2022) explored the possibility of using GenAI to provide code explanations to students. However, the effectiveness and usefulness of this approach will be evaluated at a late stage of their study. Regarding assessment design, a GPT-3 powered AI text generator was used to create exam question pools for an introductory-level data science course at the postgraduate level. The questions were evaluated by human experts and were rated favourably by them relating to the effectiveness of questions in meeting learning outcomes (Bhat et al. 2022).

AI text generators may also have the potential to help simplify the research process for academic researchers as certain types of work, such as extracting data and synthesising literature may be able to be conducted more quickly and with less (computational) effort. Researchers can, thus, devote more of their efforts to more advanced-level cognitive activities, such as data analysis and evaluation (O'Dea and O'Dea 2023b).

Challenges

There are many ethical concerns about GenAI, such as data privacy and security, copyright and intelligential property infringement, inequitable access and technological reediness (Kasneci et al. 2023; Yan et al. 2023). Among them, academic integrity is associated most closely with university students, and considered one of the key concerns raised against GenAI within the higher education context. This is because many GenAI tools are free of charge and easy to use. Users can access it through a web front or their mobile phone. In addition, it appears that the responses produced by GenAI cannot be detected easily by plagiarism checkers and AI detectors (Sullivan, Kelly, and McLaughlan 2023). And finally, ChatGPT seems to be particularly effective in producing academic writing for the social sciences and humanities disciplines, and also coding using various programme languages. It has been reported that the cases of academic misconduct have been on the rise since the appearance of ChatGPT at the end of 2022 (Tindle et al. 2023). For example, among the 1000 students surveyed recently, 48 per cent admitted that they used ChatGPT to help them with at-home tests or quizzes, and 53 per cent admitted that they used the tool to write an essay (study.com 2023). Some students also confessed to using ChatGPT to write their end-of-term module assignments (Wild 2023). The tool has also demonstrated its supreme ability in passing an MBA level exam in operations management and achieved a grade of a B or B- (Terwiesch 2023). However, it is worth pointing out that the potential and the actual misuse of GenAI may be unintentional or poorly evidenced usage, due to a lack of understanding of AI ethics among students, as well as a lack of guidance and support at the institutional level (Nakatumba-Nabende, Suuna, and Bainomugisha 2023).

The special issue includes five original research papers, using either qualitative methods, quantitative methods or mixed methods. The papers focus on the higher education sector in different country contexts, such as the UK, mainland China, Canada, Finland, Australia and Nepal. Based on their research focus, the papers are grouped into the following categories: an overview of learning and teaching; particular aspects of learning and teaching and social support and phycological well-being.

An overview of learning and teaching

Two papers emphasise understanding the potential challenges and benefits of GenAI tools in learning and teaching. One is within a single country context (e.g. China), and one is a comparative study between the UK and Nepal. Yang and colleagues (2024) emphasise student perceptions and the use of a GPT-powered chatbot (the name of the bot is unknown). The findings indicate that students and staff are positive towards the benefits of the tool (e.g. as a research partner, a brainstorming tool and an email content producer). However, concerns are raised about dishonesty and misuse of the chatbot. The authors also highlight the unexpected emotional support offered to students through their conversations with the chatbot. As with Yang and colleagues, Budhathoki and colleagues (2024) explore the adoption of ChatGPT from student perspectives. Using the Unified Theory of Acceptance and Use of Technology (UTAUT) as the theoretical foundation, the study indicates that three elements, namely performance expectancy, effort expectancy and social influence significantly impact the adoption intention for students at both countries. However, anxiety seems to have more impact on UK students than on Nepal students.

Particular aspects of learning and teaching

Adopting a process-oriented approach, Nguyen and colleagues (2024) examine the effectiveness and contribution of ChatGPT as a writing assistant to doctoral students during their academic writing process. The research was carried out with ten participants from Finland and New Zealand. The findings indicate that ChatGPT has a potential to become a powerful and effective aiding tool in academic writing. However, to achieve such effectiveness, students need to have strategic ideas and understand how to carry out clear and concise communications with ChatGPT. Can ChatGPT enhance critical thinking skills of postgraduate students? The study by Essien and colleagues (2024) answers this question. Using a mixed method and employing a sample of 107 participants studying in UK institutions, the authors suggest that the GenAI tool seems to be effective in developing skills mainly at the lower levels of Bloom’s taxonomy, such as remember, understand and apply. The tool is not so effective in developing higher level thinking skills, such as evaluate and create.

Social support and phycological well-being

In addition to its impact on learning and teaching, the following paper explores the use of ChatGPT from the perspective of social support and phycological well-being. Crawford and colleagues (2024) carried out a survey with 387 university students in Australia. The findings reveal that in the age of AI era, social support from their peers and others is still playing a key role in university students’ sense of belong. Even though student usage of GenAI tools is on the rise, and some students may have already developed some levels of social championship with AI, these GenAI tools can potentially increase the level of loneliness if AI is used as a complete substitute for human social contact.

Reflections on this special issue

This special issue addresses a topic of global importance for the higher education sector and the academia. In our view, with the rapid development of AI technologies, universities and their key shareholders, such as managers, academic tutors, support staff and students should be encouraged to incorporate such technology in their digital transformation policy, curriculum design, and learning and teaching, regardless of country context and culture context.

Our reflections are as follows. First, the papers have covered a wide range of country contexts and will help readers gain a good understanding of the potential opportunities, benefits, challenges and costs of integrating GenAI in higher education teaching and learning from global perspectives. The impact of GenAI has been explored from different aspects, such as academic integrity, critical thinking skills, AI literacy skills and social well-being. Nevertheless, most papers have focused on one particular AI text generator – ChatGPT. This is not surprising given the popularity and the widespread use of the tool. However, there are other popular open sourced GPT-powered text generators, such as Google Gemini, Bing and Claude. Empirical research on the adoption of these tools may help further strengthen the evidence on the possible contribution of GenAI to learning and teaching, assessments and student support. Second, little attention has been given to polices at the institutional level. This could be because research on GenAI in higher education is still at its infancy level, and university management is in the process of getting to know the technology (Dwivedi et al. 2023). It is however critical to explore GenAI in higher education from the university strategic level, to ensure that the adoption is aligned with the mission, vision and strategic goals of the institution and to provide explicit rules for staff and students to follow, and the training they need. Finally, the special issue only includes a small number of papers, further research is needed to explore and answer the following important questions. For example, what does ethical use of GenAI look like? How can GenAI help streamline or automate some of the key business processes in university, in addition to learning and teaching? and also what is the type of training academic and support staff need in supporting the ethical and equitable use of GenAI among students?

On the whole, this special issue has shed light on potential opportunities the benefits of GenAI to higher education globally. However, it is still too early to say whether GenAI is or will be the paradigm shift for higher education, in other words, whether GenAI adoption will drive fundamental revolution to learning, teaching and student support, even though it is showing a great potential, and there have been increased interests in using the tool.

Disclosure statement

No potential conflict of interest was reported by the author(s).