Addressing the elephant in the room: engaging students in ChatGPT conversations on assessments

ABSTRACT

The development of technology presents opportunities and challenges for the education system. This study investigates the integration of ChatGPT into higher education, focusing on tourism studies. Using a duoethnography approach, the study explores the experiences of two tourism educators who incorporate ChatGPT into their pedagogy and assessment methods. Results reveal that acknowledging students’ mixed responses towards ChatGPT encourages open discussions, while collaborative efforts in creating and assessing AI-driven assignments significantly enhance the learning process. Moreover, encouraging students to utilise ChatGPT responsibly promotes a sense of stewardship and ethical awareness, collectively enhancing educational experiences. The study presents key propositions for the effective incorporation of ChatGPT into higher education, along with practical implications for educators to leverage the potential of generative AI to enrich student learning experiences. It highlights the importance of adapting technological advancements while upholding ethical and pedagogical standards, offering insights into the effective use of AI in academic contexts.

KEYWORDS:

• ChatGPT
• generative artificial intelligence
• tourism
• duoethnography
• teaching approaches

Introduction

The advent of generative artificial intelligence (GAI) has ignited significant discussions within higher education, examining critical issues such as the evolving purpose of education, the integrity and reliability of student assessments, and the future trajectory of universities (Ivanov & Soliman, 2023). Central to these discussions is ChatGPT, a state-of-the-art language model launched by OpenAI in November 2022, designed to generate human-like text responses. The combination of an easy-to-use interface and a comprehensive range of functions offered by ChatGPT, along with its proficiency in effectively handling various types of inquiries and generating diversified content, distinguishes it as a significant breakthrough in technological innovation.

The robustness and versatility of GAI, specifically ChatGPT, have become increasingly relevant in educational contexts (Cai et al., 2023). Recent studies have illuminated the multifaceted role that ChatGPT can assume as a virtual teaching and learning assistant. Notably, it excels in offering prompt feedback to students on specific tasks (Elbanna & Armstrong, 2024; Ivanov & Soliman, 2023), including but not limited to clarifying concepts and aiding with assessment tasks. The incorporation of ChatGPT in education holds significant potential for facilitating adaptive, personalised learning (Benuyenah, 2023). However, this adoption is not without its challenges. Concerns regarding the quality and accuracy of AI-generated content necessitate rigorous oversight to ensure that students receive reliable and up-to-date information (Dalgıç et al., 2024). Additionally, ethical considerations surrounding academic integrity, data privacy, and the potential for AI to perpetuate biases must be carefully addressed to foster an ethical and inclusive learning environment (Malik et al., 2023).

Despite the impending transformations in the higher education sector instigated by GAI, research on the application of GAI in higher education lags behind. While this emerging body of literature has investigated the general application of GAI in higher education and presented two opposing views on this novel approach (Ali & OpenAI Inc, 2023; Alshater, 2023; Benuyenah, 2023; Elbanna & Armstrong, 2024), some crucial areas have been overlooked. Notably, there has been limited scholarly inquiry into educators’ actual experiences integrating GAI in teaching and assessment Le and Metzger (2024). Specifically, there is an apparent lack of in-depth examination concerning the pivotal role of effective teacher leadership in guiding students towards responsible and ethical utilisation of ChatGPT. This topic is often neglected or deemed taboo in educational discourse, representing a noteworthy research gap. Furthermore, the untapped potential of ChatGPT in shaping innovative and authentic assessment methods still needs exploration (Crawford et al., 2023; Kolade et al., 2024).

This article endeavours to address these research gaps, offering valuable insights into the evolving landscape of tourism education in the AI era. The tourism industry’s escalating adoption of AI across diverse applications, from productivity optimisation to content generation and experience design (Dogru et al., 2023), underscores the importance of endowing our students with digital proficiency aligned with industry practices. This article aims to address the elephant in the room by facilitating the incorporation of Chat GPT into our teaching and assessment design. Specifically, this duoethnography presents our reflection and experience as we collaborated with students to explore the impact of AI on assessment. The outcomes are expected to initiate an open dialogue on the incorporation of AI in tourism education and assessment design, thereby fostering a more future-ready learning environment in the current academic discourse.

Literature review

Integration of GAI in higher education

The integration of GAI tools such as ChatGPT in higher education presents several possibilities for personalised learning and assessment. By functioning as a writing assistant, providing feedback tailored to students’ needs, and generating examples that cater to their understanding levels, ChatGPT can adapt to individual learning styles (Elbanna & Armstrong, 2024). This approach aligns with the principles of adaptive learning, which emphasises catering to diverse learning needs (Benuyenah, 2023). ChatGPT can also serve as a valuable resource for students struggling with technical subjects. By clarifying complex terminology, providing clear instructions, and offering academic guidance, ChatGPT can act as a virtual learning assistant (Ali & OpenAI Inc, 2023; Alshater, 2023). Further, the creative content generation capabilities of ChatGPT offer another avenue for enriching the learning experience. By generating text, videos, and images (Elbanna & Armstrong, 2024; Kolade et al., 2024), ChatGPT can foster deeper engagement with the learning materials. This, in turn, can lead to a more effective expression of ideas and a more dynamic learning environment.

In the context of AI integration within assessments, ChatGPT presents versatile possibilities encompassing essay outline generation, grammar checks, content suggestions, and gap-filling (Crawford et al., 2023; Kolade et al., 2024). When seamlessly integrated with thoughtfully designed assessment strategies and curricula, this collaborative approach can empower students to engage in more advanced and critical forms of learning (Elbanna & Armstrong, 2024; Ivanov & Soliman, 2023). This transformation in focus liberates students from unnecessary anxieties related to reference accuracy and punctuation, affording them the freedom to delve deeper into the substantive content of their courses (Benuyenah, 2023; Crawford et al., 2023). In a recent case study, Eager and Brunton (2023) indicated that well-crafted prompts and instructional texts could significantly enhance student engagement, interaction, and learning experiences in higher education, thus broadening the scope of a student’s information access and stimulating advanced research skills. Nevertheless, research on the integration of GAI by educators remains limited.

From another perspective, advocates of AI integration assert the educators’ responsibility to impart AI-related skills to students, aligning with evolving workforce demands (Anders, 2022; Elbanna & Armstrong, 2024). They emphasise the necessity of equipping students with AI tools and techniques, given the escalating prevalence of AI technologies across industries (Alshater, 2023; Kim, 2022). Consequently, proponents argue that prohibiting AI in education may disadvantage graduates competing with AI-proficient peers in the job market. This perspective underscores the overarching objective of higher education, which is enhancing employability (Ivanov & Soliman, 2023; Kolade et al., 2024).

The opponents of AI in higher education, however, question the authenticity of learning, further worrying that the use of AI will jeopardise the integrity of the assessment. While ChatGPT can undoubtedly be a valuable tool for generating ideas and structuring students’ writing, it is essential to use it parallel with students’ analytical thinking, assessment skills, and research abilities (Alshater, 2023; Crawford et al., 2023). It is crucial to recognise the potential risks associated with overreliance on ChatGPT and to implement measures that both prevent plagiarism and ensure the accuracy and relevance of the content incorporated into students’ written assignments (Alshater, 2023; Elbanna & Armstrong, 2024; Ivanov & Soliman, 2023). For instance, the primary representative body for universities, Universities Australia, has responded by emphasising the efforts of Australian universities to revise their integrity guidelines and modify assessments to tackle the potential for students to engage in dishonest practices through the use of advanced AI (Panagopoulos, 2023). In addition, some universities in France, Hong Kong and India have banned students from using the tool (Ivanov & Soliman, 2023).

Regardless of the views, the fast-changing landscape of GAI poses a substantial challenge for the education sector to formulate guidelines that keep up pace with the latest developments and capabilities of these tools. This is reflected in a progressive shift from banning AI to ethical training in education policy (Skavronskaya et al., 2023). Consequently, some scholars underscore the crucial need to impart students with the responsible and ethical application of GAI tools in their learning and call for more research on the ethical use of AI in education (Nguyen et al., 2023; Skavronskaya et al., 2023).

Application of GAI in tourism education

Recently, educational institutions have started acknowledging the paramount importance of AI literacy in response to AI’s growing integration into diverse industries (Benuyenah, 2023; Ivanov & Soliman, 2023), including tourism, event, and hospitality sectors (Skavronskaya et al., 2023). Existing research has outlined a wide array of GAI applications in tourism and hospitality, encompassing operations, marketing, destination management, accounting and finance, strategic management, and beyond (Dogru et al., 2023). Therefore, incorporating ChatGPT into tourism education represents a pioneering step towards redefining pedagogical approaches within this dynamic field.

Despite its early stages in higher education (Kolade et al., 2024), the technology’s capacity to provide on-demand assistance and resources extends learning beyond the boundaries of traditional classroom settings, enabling students to explore and engage concepts at their own pace. Students can engage with diverse perspectives of tourism through ChatGPT, fostering a more active approach to analytical thinking, problem-solving, and logical reasoning by employing their essential cognitive skills (Dalgıç et al., 2024). The support ChatGPT offers to educators in tourism is equally noteworthy, providing interactive, real-world scenario-based exercises and content that cater to students’ interests and career aspirations (Dogru et al., 2023). This shift in focus is crucial in a field that is rapidly evolving, where tourism educators must constantly update and adapt their teaching materials to reflect the latest industry trends and practices.

Despite the relevance and significance of GAI applications, there is a lacuna in research probing into the perspectives and experiences of tourism educators in the integration of GAI into teaching and assessment practices. Many existing tourism education studies consist largely of conceptual and opinion papers. For instance, Skavronskaya et al. (2023) address the issue of academic integrity concerning AI in tourism education, while a recent study by Le and Metzger (2024) outlines a set of recommendations for tourism assessment design in tourism. While these studies provide valuable insights into the integration of AI in tourism education, a significant gap in empirical investigation persists. An even more pronounced gap is evident in the lack of engagement with students in conversations about responsible and ethical GAI applications in assessment (Elbanna & Armstrong, 2024; Kolade et al., 2024; Malik et al., 2023).

Methods

Duoethnography is a subtype of analytical autoethnography utilising a collaborative, reflective, and dialogic approach (Norris & Sawyer, 2012). The process involves two participants working in tandem to examine their autobiographical experiences, aiming to gain profound insights into their perspectives on the subject under investigation (Norris & Sawyer, 2012). In a duoethnography study, the researchers and their dialogues become the site of research (Breault, 2016). There are no prescriptive rules about the method; rather, it is guided by key tenets that provide an outline for this unique form of inquiry (Breault, 2016; Norris & Sawyer, 2012). These tenets centre around the reciprocity within the dialogues, through which new meanings can emerge (Breault, 2016). The trustworthiness of the method is anchored in self-reflexivity, while the rigour of the collaborative inquiry is made explicit in the study (Breault, 2016). The method has proven instrumental in education research for interrogating values, beliefs, perceptions, and pedagogical underpinnings of learning and teaching practices (Banegas & Gerlach, 2021; Snipes & LePeau, 2017). Duoethnograpy has also been applied in tourism research, investigating authenticity (Tan & Teoh, 2019), transformative travel experience (Pung et al., 2020), and tourism fieldwork (Manfreda et al., 2023). The present duoethnography study is based on the reflections and dialogues of two tourism educators on their experience addressing the elephant (i.e., engaging students in conversations about the utilisation of GAI for assessment) in the classroom.

One of the two courses investigated was a final year capstone course for the tourism major titled “Tourism Futures”. The course places a strong emphasis on equipping soon-to-be graduates with the skills required to address future trends and challenges in the tourism industry. The activity used in this reflection was based on an individual assessment where students assumed a stakeholder role and prepared a report to analyse opportunities and challenges for a sustainable tourism future arising from COVID-19 recovery. The activity involved a demonstration of using ChatGPT to create the report and students subsequently marking the AI-generated report in a workshop session. The marking criteria included evaluation of the introduction, content knowledge, depth of research, quality of recommendations, and writing.

The other course was a second-year course, “Sport Event and Facility Planning”, which explores both practical and theoretical aspects of event and legacy management. The course assessment comprised three components, including the drafting of a lobbying letter to a local state member of Parliament in regional Queensland, critiquing Brisbane’s 2032 Olympic Games bid. Students were required to research and articulate the potential drawbacks and opportunity costs of hosting the Games, considering social, environmental, and economic impacts. Two weeks prior to the deadline, the instructor integrated a workshop session involving ChatGPT, where students used the tool to generate a lobbying letter based on the assessment instructions. Students then evaluated the letter against the marking criteria, assessing its purpose, content knowledge, sources of supporting evidence, and the overall structure and quality of writing.

The data in this study aligns with the approach taken by past duoethnography in education (Snipes & LePeau, 2017) and tourism (Pung et al., 2020), which entails personal reflective essays and transcripts of conversations. Both educators independently recorded their reflections on the respective activity, detailing students’ reactions, the role of educators in facilitating responsible utilisation of GAI, and the opportunity to engage GAI in authentic assessment design. This is followed by three rounds of conversation to unpack their experiences from August to October 2023. The conversations were transcribed and coded thematically alongside the written reflections. The themes that emerged from the data are detailed in the subsequent section. The names of the authors have been replaced by pseudonyms for peer review purposes.

Results

Acknowledging the elephant in the room: addressing the discomfort and thrill

The announcement of the workshop activities on ChatGPT was met with a range of emotions in the classroom, from uneasiness to excitement. Tina reflected: ”Initially, when I shared the idea of incorporating Chat GPT into our upcoming workshop, I observed that students were somewhat hesitant and uncomfortable. It seemed as if they considered this topic almost taboo within the academic setting. However, as I went on to explain the rationale behind this exercise and its potential benefits, a noticeable transformation occurred. Their initial reluctance gave way to a growing sense of excitement and curiosity.”

Maia made a similar observation: “When I announced that we are going to use Chat GPT, I sensed excitement in the class by looking at students’ facial expressions. It almost felt like teenagers doing something they were not supposed to do. Like a taboo.” In Maia’s class, students mentioned that they had been advised to stay away from ChatGPT in other courses.

Both educators noticed a significant surge in student engagement and participation in those workshop sessions. Tina attributed this to the activity “reflecting their genuine enthusiasm and interest”, highlighting the importance of engaging students in conversations about GAI.

Let’s touch the elephant together: creating and assessing AI-Generated assignment

The activity began with generating the assignment using ChatGPT. Neither assessment guide had been deliberately AI-proofread, as they were created before the widespread introduction of GAI. As Maia copied the assessment instructions into ChatGPT, she noticed this in class: ”We could see on the screen that a report was being generated within seconds. Everyone was in shock. What took these students weeks to try and get their ideas down on paper was now being materialised in just seconds. Everyone’s eyes were glued to the screen as the words were being punched out by Chat GPT.”

Tina, observing a mix of curiosity, disbelief, and uncertainty among her students, reflected on the ethical implications of GAI integration in learning and teaching practices. She noted, ”I think what struck them was the fact that, as their instructor, who should typically discourage the use of such tools due to the prevailing perception that ChatGPT was associated with cheating or unethical behaviour, I was not only permitting its use but actively endorsing it.”

Students were then instructed to evaluate the AI-generated assignment using the actual marking criteria. Both educators noticed a high, intense and extraordinary level of concentration in their classes as students curiously or sceptically undertook the activity. The marking activities revealed the possibility of achieving a passing grade using ChatGPT in assessment with minimal manipulation. However, students were highly critical in their marking and identified several shortcomings of the AI-generated assignments. These included the content needing to be more generic, making little reference to course content, lacking supporting evidence, and having a low level of critical analysis.

From the instructors’ perspectives, Tina observed that while ChatGPT had produced an assignment that adhered to the guidelines, maintained a concise format, and demonstrated grammatical correctness, it relied on verbatim words from the instructions without delving deeper into their meanings and justifications. Maia concurred and noted that assessment instructions were unpacked and explained at greater length in class. She commented: ”The expectation of assessment is often negotiated between the educator and the learners through their interactions in class … There are subjective influences in our communication to students in terms of what they should be including in the report and how we will be marking them. These nuances are not captured by Chat GPT solely relying on the assessment guide.”

Co-creation with the elephant: empowering students to responsibly utilise ChatGPT

Following the activities, Tina and Maia engaged students in constructive conversations about their experiences using ChatGPT for assessment and key lessons learned. To fully harness the benefits of GAI, students are required to comprehend the assessment requirements, marking criteria and course content. Tina reflected on how this activity has achieved its learning objectives, which included helping students to better understand the assessment instructions, how they would be graded, and the relevance of the assessment to what they were learning in the course. Sharing a similar sentiment, Maia noted, ”They [students] are more consciously aware of the standards that are expected in order to do well in this assessment. It was their words of the lack of depth and connection with the course content.” Overall, students commented that the exercise had given them some clarity about assessment and reduced their anxiety, as they felt that they now had something to start with.

Through the activities, the instructors were also able to co-create the expectation of the assessment with students, discussing what needs to be included in the assessment and the expected level of standard, especially with the imminent proliferation of GAI integration. Maia reported, ”We agree that the advent of AI will inevitably raise the bar higher when it comes to expectations and marking, and students need to put in more effort.‘ Likewise, Tina stated, ’We collectively reached an agreement that relying solely on ChatGPT would result in a minimum passing effort. … It became evident that students should take the lead.”

Above all, students have been empowered to co-create their assignments responsibly with inputs from GAI, sanctioned by the instructors. Tina remarked, ”Students felt empowered, as they had just marked the assessment and now held the ability to propose improvements.‘ Maia declared, ’I explicitly told my students to use Chat GPT by all means. Use the tool to generate preliminary ideas, but they are expected to do further research on these ideas, deepen the analysis and relate to the course content.”

Discussion

The emergence of ChatGPT has sparked substantial discussions pertaining to the prospective trajectory within higher education. A range of contrasting viewpoints exists regarding its suitability (Elbanna & Armstrong, 2024; Kolade et al., 2024). In light of the research objective to investigate the use of ChatGPT in the teaching and assessment methods within tourism education, it becomes crucial to address how these technological innovations can be employed while maintaining ethical and educational standards. Based on the findings of this study, several propositions have been developed. These are intended to assist educators and academic institutions in overcoming the challenges and leveraging the opportunities that arise with the introduction of generative AI in higher education. Table 1 provides an overview of these essential propositions and their respective implications.

Table 1. Key propositions and implications for higher education.

Propositions	Implications for Higher Education
Addressing Students’ Discomfort and Excitement	Acknowledging students’ mixed feelings towards ChatGPT can lead to a more open and productive dialogue on its integration in educational settings.
Collaborative Creation and Assessment	Collaborative creation and assessment of AI-generated assignments can enhance the students’ learning experience and ensure a deeper understanding.
Empowering Responsible Use	Empowering students to use ChatGPT responsibly fosters a sense of ownership and encourages ethical considerations in its application.

Numerous academics have expressed concerns over the efficacy of its implementation within educational settings, predominantly highlighting concerns about copyright violations, authentication, and preserving academic integrity (Crawford et al., 2023; Ivanov & Soliman, 2023; Malik et al., 2023). Rather than adopting an exclusively critical perspective towards ChatGPT, such as mitigating or preventing the use of AI in assessment (Ivanov & Soliman, 2023), this study advocates for a dialogical learning and teaching approach underpinned by the pedagogy of discomfort (Boler, 1999). This duoethnography study highlights the role of educators in moving students outside of their comfort zone and facing the elephant (i.e., ChatGPT) in the room. The feeling of discomfort reported in this study indicates that students’ underlying assumptions and beliefs were being challenged, which could create openings for individual transformation and, in this context, responsible and ethical GAI integration in assessment and learning more broadly.

The key propositions further emphasise being actively engaged with students in conversations about GAI, with the primary objective of empowering them to apply this technology responsibly. This can be achieved by openly collaborating with students to explore the potential and limitations of AI-generated assignments in the classroom, thereby instilling a deeper understanding of the assessment tasks and expectations through this process. The learner-centric approach invites students to examine their positions on the use of AI in assessment, imparting not only subject knowledge but also reflective and self-regulation skills (Nguyen et al., 2023). This approach extends beyond the typical academic integrity declaration, where students ethically disclose if the assignment was created with the assistance of AI (Skavronskaya et al., 2023). Furthermore, these candour conversations offer an opportunity for educators to address any concerns students may have regarding the ethical and responsible application of AI.

The use of GAI will continue to evolve in higher education, and human-technology integration will be increasingly important in future workplaces. Central to these discussions is a key question: How can we accurately evaluate students’ learning outcomes when AI contributes to assignment creation? This dilemma necessitates a re-evaluation of traditional assessment methods, prompting educators and institutions to explore innovative approaches (Alshater, 2023; Crawford et al., 2023). This includes moving away from simple recall of facts to assessment tasks that promote critical and analytical thinking and responsible AI use. Digital literacy, interdisciplinary learning, and problem-solving skills should also be nurtured, ensuring that students can leverage ChatGPT effectively while developing essential competencies for academic and professional success (Anders, 2022; Elbanna & Armstrong, 2024; Ivanov & Soliman, 2023).

In light of these considerations, tourism educators are tasked with a critical responsibility: to re-evaluate and redefine the foundational skills and knowledge that students need to thrive in the future landscape of tourism (Dalgıç et al., 2024). This entails not only integrating ChatGPT and similar technologies into the curriculum but also developing strategies to mitigate their limitations. Le and Metzger (2024) have suggested ways to balance the influence of AI tools in tourism education, including prioritising assessment tasks that require students to gather and critically analyse real-world data such as industry reports. Another opportunity for tourism education includes incorporating decision-making simulations that can equip students with the skills to evaluate their choices and predict outcomes based on complex information and multiple factors (Pratt & Hahn, 2016).

Apart from assessments that focus on research and analysis that go beyond the capabilities of AI tools, we also recommend designing assessments that encourage students to collaborate with AI to solve problems or discuss complex topics. One example would be multimodal scenario planning of sustainable tourism development with the assistance of AI where students are required to report the prompts and data used to generate the scenarios, evaluate the AI-generated scenarios, and present the outcomes in a multimodal assessment format. This format could include infographics, interactive oral presentations, or even video productions, rather than the traditional text-based assignment. By introducing innovative human-AI collaboration in assessment, universities could instil in students the readiness to be agile and adaptable to the changing future trends with responsible AI use, instead of shying away from technological advancements (Benuyenah, 2023). This approach enables educational institutions to fully leverage the potential of AI to enhance tourism education, equipping students with the information literacy, digital proficiency, and ethical awareness needed to navigate the challenges and opportunities of the industry.

Conclusion and limitations

While GAI holds the potential to support students in their educational endeavours, it is crucial to recognise that it should not be considered a substitute for traditional learning methods. Instead, GAI should be seen as a supplementary tool that offers an alternative pathway for acquiring knowledge (Alshater, 2023). Educators seeking to assist their students in this transition may find it appropriate to utilise ChatGPT to provide constructive forms of assistance. This study demonstrated the value of engaging students in discussions about using ChatGPT in assessment. Specifically, the class activities in focus offered alternative perspectives on preparing for assessments in the AI era. The learning outcomes include elevating students’ submissions by enhancing their understanding of assessment criteria, marking rubrics, and academic integrity. ChatGPT has the potential to help students identify overlooked areas and provide additional study recommendations while addressing the tool in the classroom instils confidence in the validity of utilising this tool responsibly to enhance students’ assessments.

While this research explores the potential of using ChatGPT in teaching and assessment methods in tourism education, it is crucial to acknowledge limitations that call for future research. The current focus solely on instructors’ perspectives in two specific courses might restrict the generalisability of the study findings. To gain a comprehensive understanding, future research should involve a wider range of educators across diverse tourism programs. Furthermore, including students’ perspectives would offer further insights. One area of focus could be measuring the impact of employing ChatGPT on student learning outcomes, particularly in fostering critical thinking and responsible AI use. Finally, research is warranted on the development of standardised practices and frameworks for ethical and effective integration of ChatGPT in tourism education course content and assessments.

Disclosure statement

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

Additional information

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.