Abstract
Keywords
Introduction
Interdisciplinary qualitative research (IQR) integrates diverse dimensions and offers multiple conceptual and methodological approaches to explore a complex social phenomenon and human behavior across disciplines (Sabo et al., 2023; Vaughan, 2005; Wolgemuth & Jordan, 2023). IQR methodologists often identify their subjectivities as social scientists or critical social researchers, navigating lived experiences, emotions, and perceptions in different social phenomena (Trussell et al., 2017). Thus, they encounter complexities and risks among collaborators whose subjectivities diverge across time, identity, and space (Anders & Lester, 2015; Bark et al., 2016; Bromham et al., 2016; Trussell et al., 2017). Notably, Nam and his colleagues emphasized the complexity of subject-object relationships in a broader global context, primarily when co-authors represent transnational, intercultural, bilingual, multilingual, or multicultural identities (Nam et al., 2024b, 2025a, 2025b). Put simply, subjectivity involves various aspects and qualities of social scientists in qualitative research, such as reflexivity, positionalities, and social representations. It helps readers understand authors’ insider, outsider, or in-between roles, as well as potential biases toward different individuals and social groups (Merriam & Tisdell, 2016). Additionally, the subject-object relationship relates to how reciprocity enhances generalizability by fostering mutual rapport and trust between researchers and participants (Denzin & Lincoln, 2011).
Despite the many forms of qualitative methodologies and methods, it is challenging to apply or combine different approaches when collaborators have divergent subjectivities, disciplines, knowledge-building aims, research designs, and proposal questions (Nam et al., 2024b; Sabo et al., 2023; Trussell et al., 2017; Vaughan, 2005; Wolgemuth & Jordan, 2023). Trussell et al. (2017) discussed the complexities and risks of interdisciplinary qualitative research, especially when co-authors come from different disciplines and hold varying views on knowledge dissemination. They may encounter emotional conflicts during team building and when negotiating internal and external pressures, such as vulnerability and emotionality within the research team and concerns about the presentation, legitimacy, and academic value of the research (Trussell et al., 2017). Therefore, IQR methodologists face power dynamics related to academic identities when forming research teams. They consider their different positions and methodological expertise, as well as mainstream perspectives in senior-junior academic relationships (Nam et al., 2024b, 2025a, 2025b). These IQR methodologists also deal with various issues, including conflicts of interest, authorship disagreements, and funding sources (Anders & Lester, 2015; Bromham et al., 2016). Notably, social scientific research paradigms and trends have gradually shifted toward the informatization of education, data-driven policy models, the knowledge-based economy, and human capital in the current era of post-digital capitalism and the fourth industrial revolution (4IR), along with the rapid development of generative artificial intelligence (AI) models (Gan & Bai, 2023; Nam & Bai, 2023; Peters et al., 2022, 2024).
We are two co-authors representing our subjectivities as IQR methodologists at the intersection of science, technology, engineering, arts, and mathematics (STEAM), business, management, and economics (BME), and educational leadership and policy studies (ELPS), all within the broader field of higher education research and development (HERD). The first author holds bachelor’s, master’s, and Ph.D. degrees in STEAM-related areas, specifically art design, animation production technology (APT), and digital communication and media arts (DCMA). The second author has bachelor’s, master’s, and Ph.D. degrees in fields related to BME and ELPS. As qualitative methodologists, university faculty members, academic editors, ad hoc reviewers, and business entrepreneurship trainers, we have routinely pondered crucial questions about IQM and related issues concerning the impacts of generative AI models on our various fields.
We share concerns about issues affecting our communities, including potential research and business ethics issues, pedagogical challenges, and human capital crises that generative AI models could raise. Therefore, we aim to capture our subjectivities through a meta-autoethnography of metadiscourse, offering methodological implications for IQR in the context of generative AI models. We pose critical questions: In what ways do IQR methodologists offer plausible implications for the global public and intellectual society? More specifically, what strategies can different co-authors’ subjectivities and their conceptual clarity use to strengthen particular research designs, corpus generations, the ideas of trustworthiness and bias statement, and rationale for specific research scope in the given naturalist inquiry and its existing bodies of literature?
Overall, we aim to demonstrate research procedures by dividing into five primary sections: (a) existing qualitative literature on generative AI models in divergent fields; (b) philosophical assumptions and interpretive frameworks; (c) designing a meta-autoethnography of metadiscourse on generative AI models; (d) meta-autoethnographic writings and potential themes in the metadiscourse; and (e) concluding remarks: methodological implications for IQR on generative AI models. To this end, we will present specific procedures; elaborate on methodological rigor; address issues of relevance and applicability within applied social scientific research; and articulate the contextual engagement, focus, clarity, and rationale for conducting this IQR.
Existing Qualitative Literature on Generative AI Models in Divergent Fields
A Brief Overview of the Development of AI Research
When conducting social scientific research, especially in qualitative inquiry, it is crucial to identify the scope of research topics and gaps (Egbert & Sanden, 2015). As our primary research topic concerns generative AI models across our differing fields, we offer a brief overview of the development of AI research. Briefly, generative AI models are computing systems, through which humans use their intellectual and cognitive abilities to ask questions and receive responses; these models are currently known as text-to-text and text-to-video models. It shapes how people can learn, self-correct, adapt, and synthesize data for complex tasks through language and visual support (Liu et al., 2024; Mogavi et al., 2024; Salas-Pilco & Yang, 2022).
Indeed, the mainstream focus of AI research has traditionally been dominated by scholars from science, technology, engineering, and mathematics (STEM) fields, particularly in areas like computer science, information and communication technology (ICT), software engineering, applied statistics, and data science (Chiu et al., 2025; Li et al., 2020, 2022; Nam & Bai, 2023). Recently, new academic paradigms and innovative curricula have introduced the “STEM+ the Arts” (STEAM) approach (Li et al., 2020, p. 8). Conventional arts scholars have contributed to practical applications through “technological pedagogical and content knowledge” (TPACK) and assist “teachers and learners understand content-based learning and pedagogical strategies to build confidence in using new technology” (Gan & Bai, 2023, p. 1517).
From a social network perspective, many STEM/STEAM researchers have begun collaborating with social scientists from various disciplines to explore how generative AI models influence human intelligence. This includes fields such as BME, ELPS, and broader HERD (Hu & Sthal, 2023; Nam et al., 2025a; Nam & Bai, 2023; Rousell & Sinclair, 2025). Consequently, AI research has become an interdisciplinary domain within the scientific knowledge society, incorporating qualitative, quantitative, and mixed-methods approaches. They have represented diverse viewpoints, including optimism, pessimism, and skepticism about generative AI (Nam et al., 2025a; Nam & Bai, 2023).
The Emergence of Qualitative Research on Generative AI Models and Its Gaps
In reviewing recent qualitative literature, scholars from various disciplines have helped expand knowledge in their fields. For example, a series of systematic review papers analyzed research trends in STEM over the past twenty years. They summarized key topics, publication patterns by country, and emerging trends and priorities, with increasing focus on AI research (Chiu et al., 2025; Li et al., 2020, 2022). According to Li et al. (2020, 2022), STEM research has mostly been led by scholars from higher education institutions (HEIs) in the United States, followed by Canada, Australia, and the United Kingdom. However, arts scholars, social scientists, and critical pedagogues have described themselves as “STEAM” researchers, embraced interdisciplinary education, and published influential “empirical research publications in STEM education” (Li et al., 2022, pp. 2–3). Similarly, Chiu et al. (2025) pointed out that computer scientists have traditionally dominated AI research. Nonetheless, STEAM researchers are also integrating AI and advanced technologies such as robotics, nanotechnology, 3D printing, and biomechanics into their work (Gan & Bai, 2023). As a result, these interdisciplinary fields are increasingly seen as “AI-STEAM” (Hsu et al., 2021, p. 1).
Other discourse and reflexive narrative papers have also used textual data to present diverse empirical perspectives and deepen understanding of learner identity construction in STEM and moral dilemmas related to chatbots in HERD (e.g., Bearman et al., 2023; Hu & Stahl, 2023; Nemorin et al., 2023). Notably, some recent self-studies have addressed ethical issues associated with generative AI and/or large language models (LLMs). For instance, Chubb (2023) experimented with AI chatbots to translate arts-informed data into vignettes and discussed the opportunities and challenges of AI in qualitative data analysis. Hayes (2025) also engaged with LLMs through personal interaction with generative AI models and tried to simulate conversations using qualitative data. These scholars commonly raised concerns about research ethics and academic standards, such as data security, bias, prejudice, and offensive language, often pointing out issues that are not always accurate.
We view qualitative researchers as facing different challenges related to epistemological traditions, theoretical frameworks, and paradigms. However, the role of methodology could help bridge research gaps across divergent fields by leveraging interdisciplinary knowledge to advance applied social scientific methodologies. They focus on central arguments and endeavor to strengthen the logic by linking divergent ideas coherently (Sabo et al., 2023; Vaughan, 2005; Wolgemuth & Jordan, 2023).
We contemplate applying meta-auto ethnography into meta discourse as the primary methodological tool; meta-auto ethnography is described as “a form of qualitative meta-analysis” and involves researchers’ reassessment of “their own body of work to gain more ideas [and] more insights,” which provides an updated and renewed sense of self (Hughes & Noblit, 2017, p. 211). Furthered, metadiscourse refers to “discourse about discourse” and the “linguistic manifestation in a text to ‘bracket the discourse organization and the expressive implications of what is being said’…therefore a crucial rhetorical device for writers” (Hyland, 1999, p. 5). We conceive that both meta-auto ethnography and metadiscourse involve co-constructive reflexive narrative and interpretive work.
Philosophical Assumptions and Interpretive Frameworks
Liberal Practicality and Co-Constructivist Approaches to IQR
Qualitative researchers articulate their philosophical assumptions and interpretive frameworks as the initial stage of their research designs (Creswell, 2013). However, when it comes to IQR, these researchers need to consider the strengths, weaknesses, and limitations of their chosen approaches. Given the challenges faced by IQR methodologists and their ambiguous subjectivities, earlier social scientists adopted liberal practicality and co-constructivism as their philosophical orientations. For example, in his
From these perspectives, we believe that IQR methodologists can learn from one another through co-constructive narratives and/or collective writing projects. They can develop new research agendas on specific topics by engaging in mutual self-reflection, dyadic and interactive interviews, and follow-up communications (Hughes & Pennington, 2017; Nam et al., 2024b, 2025a, 2025b; Peters et al., 2022, 2024). These collaborators can deepen a holistic understanding of power dynamics in real-world situations. However, there are certain limitations to IQR. Numerous research topics and themes exist within similar fields. Therefore, IQR methodologists must recognize that there are specific expertise constraints and/or limitations in the data sources used for evidence (Anders & Lester, 2015; Bark et al., 2016; Trussell et al., 2017). At this point, we note that all social scientific research has scope and methodological limitations, which should motivate future scholars to continue expanding knowledge in global scholarship (Egbert & Sanden, 2015).
Pragmatist Ethics and Critical Pedagogy in the Age of Generative AI Models
As generative AI models have raised ethical dilemmas among different users, we consider underpinning our approach with pragmatist ethics and critical pedagogy. Similar to liberal practicality (see again Mills, 1959), pragmatism is a philosophical assumption and interpretive framework used in social scientific research and qualitative inquiry. This ontological and epistemological approach emphasizes the practical consequences of thoughts, situations, and actions, aiming to enhance critical thinking, problem-solving, and improve human lives in practical ways (Creswell, 2013).
Regarding the classical roots of pragmatism, William James’ (1907) work,
Israel (2014) discusses the researcher’s moral and ethical considerations regarding research integrity. Key issues include authorship and plagiarism, which involve using others’ work or ideas developed through their efforts. Therefore, academic journal editors and their boards play a vital role in promoting ethical researchers and writers, preventing ghost authorship, and ensuring authors’ substantive involvement in their research (Kim, 2023). Ethical researchers are diligent and trained through extensive engagement in ethical practices, including data collection, analysis, and reporting. They follow specific guidelines to improve their academic writing and uphold personal integrity (Creswell, 2013).
Additionally, pragmatist business maxims are closely linked to organizations’ success, financial value, and consumer behavior. Hence, a pragmatist approach to integrity in business ethics involves developing integrity through social awareness, emotional intelligence, self-awareness, and effective intercultural and interpersonal communication. As a result, organizational policies and practices reflect ideals of ethical leadership, moral conduct, state, and commitment (Maxcy, 1991). Business ethics, grounded in objectivism and corporate integrity, fosters loyalty, a set of principles, and long-term goals that promote an organization’s success (Barry & Stephens, 1998).
Concerning critical pedagogy, John Dewey’s early pragmatic works,
To date, perhaps the most well-known Generative AI models are OpenAI’s platforms, such as Chat Generative Pretrained Transformer (ChatGPT) and Sora; ChatGPT is a text-to-text model, and Sora is a text-to-video model (Liu et al., 2024; Mogavi et al., 2024). Despite the promise of generative AI models, ethical concerns and conflicts are growing among various stakeholder groups. Notably, the emergence of text-to-video models has led many intellectuals and practitioners to reassess the use of AI and its implications for the future of human intelligence because it can simulate visual and physical motions in dynamic training systems (García-Peñalvo, 2023; Kim, 2023; Kumar et al., 2024; Nam & Bai, 2023).
Today, higher education institutions (HEIs) serve as the foundation for teaching and research, supporting scholars and professionals across a wide range of fields. Hence, new technology and digital devices have become practical tools to enhance the success, enrollment, reputation, and philosophical awareness, as well as digital literacy and creative thinking, of HEIs and their students. However, educational leaders, faculty members, and students face various moral dilemmas when balancing theory and practice in educational leadership and policy studies, particularly in areas like academic services and college student development (Evans et al., 2010). From these perspectives, we reflect on how digital divides in the age of generative AI may lead to conflicts among different stakeholders in academic publishing, business entrepreneurship, teaching, and the human capital crises, both within the academic and professional labor markets. Therefore, we consider these pragmatist ethics and critical pedagogy as our fundamental philosophical and interpretive frameworks.
Designing Meta-Autoethnography of Metadiscourse on Generative AI Models
Methodological Rationales for Meta-Autoethnography of Metadiscourse
Qualitative researchers view issues of methodological relevance and applicability as fundamental to their research designs (Creswell, 2013). Therefore, we provide methodological reasons for using meta-autoethnography of metadiscourse. In general, autoethnography encompasses various forms that explores “identities, power, privileges, and penalties within one or more cultural contexts” (Hughes & Pennington, 2017, p. 7). Each author compares or contrasts their self-studies or life histories, such as autobiographical writings, through a specific theoretical lens (Hughes & Pennington, 2017). When two or more authors collaborate on similar works, it is called duoethnography and/or collaborative autoethnography, employing critical, evocative, analytical, and reflexive narrative tones to contextualize or recontextualize their own past and present. These approaches ultimately aim to foster positive social change (Anderson, 2006; Hughes & Pennington, 2017).
Despite the growing body of research exploring the role of autoethnography and discourse in generative AI models, methodological depth in IQM and applied qualitative methods remains limited. From a pragmatist perspective, Creswell (2013) argued that each researcher has the freedom to choose “methods, techniques, and procedures of research that best meet their needs and purposes,” so “worldviews will use multiple methods of data collection to best answer the research questions,” and “focus on the practical implications of the research” (p. 28). From this viewpoint, some notable examples of this collaborative approach include designing meta-ethnographies of autoethnographies and duoethnographies of ethnographies (e.g., Hughes & Noblit, 2017; Hughes & Pennington, 2017), meta-narratives into metadiscourse (e.g., Jiang & Hyland, 2025), critical discourses into meta-autoethnography of descriptive mixed-methods (e.g., Bai & Nam, 2020, 2022, 2023a, 2023b, 2024; Nam et al., 2024a), and co-constructive narratives into digital ethnography (e.g., Nam et al., 2022, 2024b).
In a nutshell, qualitative research methodologists have aimed to co-develop more practical and sophisticated methodologies and methods, thereby expanding existing methodological knowledge and its implications for broader audiences. This indicates that their approach is not restricted to specific methodological traditions, such as traditional ethnography, autoethnography, reflexive narrative, and basic qualitative methods like interviews and observations (Gobo & Molle, 2017; Hughes & Pennington, 2017; Merriam & Tisdell, 2016). Therefore, we consider integrating a meta-autoethnography into metadiscourse, as both methodologies entail reflexive practice, co-constructive narrative, and mutually joint scholarship (see again Hughes & Noblit, 2017; Hyland, 1999).
Meta-Auto Ethnography of Meta Discourse as a Co-Constructive Reflexive Methodology
In general, a co-constructive lens can support co-authors in IQR when two or more authors from different subjectivities collaborate on works committed to collective action (Hughes & Pennington, 2017; Nam et al., 2024b, 2025a, 2025b). Nam et al. (2024b) stated that this lens focuses on seeking “to generate a more in-depth empirical understanding of the co-authors’ collective interpretation of the complexities inherent in friendship and community relationships, mainly when aiming to explore confusion, uncertainty, ambiguity, and contradiction in a particular social phenomenon” (p. 3). Thus, collaborators’ reflexive turns on the subject-object relationships are significant for elaborating on particular social phenomena, thereby bolstering the quality of their reflexive turns through their subjectivities (Hughes & Pennington, 2017).
Grounded in a co-constructivist lens, we draw on meta-auto ethnography and meta discourse to inform an applied qualitative approach, aiming to articulate the different elements and functions in the development of IQR. Initially, meta-auto ethnography is part of an applied qualitative methodology that requires a systematic analysis of researchers’ prior autoethnographic work and/or qualitative lines of inquiry (Hughes & Noblit, 2017; Hughes & Pennington, 2017). This approach differs from other forms of auto ethnography, meta-analysis, and systematic review. While other forms of auto ethnography represent single, duo-, collaborative, and/or collective reflexive writings to raise critical questions about power, honor, prestige, or alienation, marginalization, and stigmatization in various fields, meta-auto ethnography focuses on authors’ layered accounts of new interpretations from their previous research.
Whilst meta-analysis and systematic review approaches deal with larger volumes of existing quantitative and qualitative literature and its hypotheses and assumptions, meta-auto ethnography focuses on “a systematic process of critical reflexive thinking and synthesis of one’s own previous auto ethnography work in order to learn from it and through it” (Hughes & Pennington, 2017, p. 20). Nonetheless, Hughes & Noblit (2017) suggested meta-auto ethnography as an exemplar, drawing on their previous qualitative literature and that of others more broadly, and offered a clearer picture of the nature of methodological implications: …our goal is to learn what authors’ lives taught them and what it evokes in us, while relating our lives and learning to each other. Through this way of relating, we come to say something about the meanings of others’ lives – and our own. The meta-ethnography of autoethnographies produces analogies that allow readers to anticipate possibilities rather than predict the meanings of lives lived in certain contexts and under certain conditions. Synthesis is driven by purpose. It is helpful to think about purpose in terms of a phenomenon of interest. Each autoethnography encompasses many possible phenomena of potential interest. Thus, the interest of the synthesiser is paramount (p. 224).
In addition to metadiscourse, discourse as a general conceptual approach fosters scholarly discussions about public concerns, cultural events, and social issues, paying close attention to both told and untold stories, while delivering messages to support ongoing social change (Bai & Nam, 2020, 2022; Fairclough, 1992; Giddens & Sutton, 2015). As a methodological approach, Fairclough (1992) emphasized the textual nature of critical discourse analysis (CDA), which explores “collaborative” or “competitive modes of interaction” among members of society (p. 206). CDA is a dialogical, dialectic, debate-driven, and competitive approach to critical social research, utilizing relational dialectics and communication to deepen understanding of various concerns, issues, events, experiences, perceptions, and conflicts within a public setting. CDA researchers often consider a variety of textual data sources. These broadly include, but are not limited to, archival records, media representations, memoirs, reflexive diaries, and oral testimonies (Fairclough, 2003; Smith et al., 2025).
Meanwhile, metadiscourse refers to the facets of textual and hermeneutic data sources that evidence both divergence and convergence in public and academic settings (Hyland, 1998, 1999). The conceptual elements of metadiscourse illustrate “those aspects of the text which explicitly refer to organisation of the discourse or the writer’s stance towards either its content or the reader. While the term is not always used in the same way…the distinction between the ideational elements of a text and its textual and expressive meanings” (Hyland, 1998, p. 438). The importance lies in collecting sufficient textual and hermeneutic data to cluster patterns, despite the number of texts not being necessarily quantified to ensure generalizability, given the nature of qualitative inquiry. However, the goal of metadiscourse is to employ a hermeneutic and textual analysis approach, organizing categories and linking terms, phrases, sentences, and passages to achieve a comprehensive understanding of the chosen argumentation (Hyland, 2017; Jiang & Hyland, 2025). The hermeneutic and textual data sources of evidence broadly include, but are not limited to, peer-reviewed publications, textbooks, public and social media sources, archives, academic presentations, and oral testimonies available on digital and online databases. Therefore, discourse analysts review the chosen textual and written datasets, categorize and select them to represent the original meanings in an academic tone (Hyland, 1998, 1999; Nam & Bai, 2023).
Handling Reflective Writings and Hermeneutic Data
We triangulated our reflective writings and hermeneutic data with various textual sources from our previous co-authored or individual publications in IQR, as well as academic and public writers. From these standpoints, we also collected our publications that discuss power differentials and issues related to the advent of AI, as well as structural problems associated with the digitalization of education and cultural movements. However, we acknowledge that our previous work had certain limitations regarding our subjectivities, which should be recontextualized or reinterpreted to offer novel insights into global scholarship. Considering our subjectivities, we perceive that our target audiences are the global public and intellectual society, rather than specific gender, racial, ethnic, or national groups. Accordingly, we position our researchers’ subjectivities within the fields of STEAM, BME, and HERD, aligning with our academic and professional roles as members of the global academy and industries.
Based on our subjectivities, philosophical assumptions, and interpretive frameworks, the inclusion criteria for written and textual readings are as follows: (a) positional statements regarding the use of AI as accepted, denied, or natural at a general level; and (b) ethical issues, conflicts, anxieties, and potential human capital crises in each domain. We considered academic publishers as our starting point, including
We also reviewed and organized articles from news magazines and archival organizations. A total of 118 sources were examined, from which 46 of the most influential and relevant were selected to develop metadiscourse; we carefully chose written and textual evidence. We considered the visibility and prominence of the topics and their potential outcomes, such as themes, patterns, categories, and/or domains (Creswell, 2013; Merriam & Tisdell, 2016). The co-constructive narrative approach allows collaborative autoethnographic or duoethnographic writers to explore their work. For example, Hughes and Pennington (2017) stated that this approach navigates “rational experiences, particularly how people collaboratively cope with the ambiguities, uncertainties, and contradictions of being friends, family, and/or intimate partners…relationships as jointly-authored, incomplete, and historically situated affairs” (p. 18). Therefore, all these sources are relevant to the liberal, practical, and pragmatic approaches, alongside our subjectivities and the issues within our shared relationships.
Among these were (a) existing scholarly literature, including our own publications; (b) editorials in academic journals such as
Overall, we specifically considered journal editorials from
Recontextualization and Trustworthiness
In the given naturalist inquiry, qualitative methodologists bolster trustworthiness by adopting data collection strategies and respecting “
Moreover, confirmability illustrates the degree to which the data report should minimize potential bias, especially given the nature of the subject-object relationships (Shenton, 2004). Accordingly, we invite three experts in IQR, who are academic editors and experienced reviewers in the fields of applied psychology, STEAM, BME, English language teaching (ELT), and corpus studies. Lincoln and Guba (1985) highlighted the role of audit trail, in which scholars outside their research support can review, check, and control potential biases. We consistently communicated with them to reduce our biases toward the current research subject-object relationships. Their personal opinions and suggestions are also partially ingrained in our current work through the reciprocity activities. More specifically, we shared our personal life stories and intentions for writing this paper, as well as our textual materials. After we completed the initial draft, they reviewed our previously published papers and personal autobiographical writings. They checked whether our voices are candid or not, including our traditionally disciplined areas, newly formulated IQM fields, our contributions to scholarship, and our ad hoc experiences as academic editors and reviewers.
Additionally, credibility refers to how data can be triangulated and what potential contents or themes can be logically categorized (Creswell, 2013). We selected credible written and textual data sources of evidence from multiple outlets. These included academic editorials, industrial publications/magazines, archives, and peer-reviewed publications, as well as our previous works and current reflective writings. There may be some concerns regarding credibility and bias in the context of a post-positivist view in qualitative research. Nevertheless, pragmatism and co-constructivism are beneficial for IQM. Creswell (2013) addressed that a postpositivist lens certainly allows the use of a “belief system” yet not always believe “in strict cause and effect, but rather recognize that all cause and effect is a probability that may or may not occur” (pp. 23–24). Thus, to recontextualize the data, we compiled, reviewed, and categorized these diverse hermeneutic data sources, acknowledging that the truth is approximate but not always entirely grasped in the given naturalist inquiry (see Creswell, 2013; Denzin & Lincoln, 2011; Lincoln & Guba, 1985). Finally, dependability explains how data collection is documented and protected from other similar works (Denzin & Lincoln, 2011). Given this, we retained the original textual meaning as much as possible, while paraphrasing them from the previous writings. We preserved the intentional meanings and their rich descriptions in the meta-autoethnography of metadiscourse (Hyland, 1998, 1999).
Meta-Auto Ethnographic Writings and Potential Themes in the Metadiscourse
Relevant Issues for Judy as a Critical Pedagogue in STEAM Education and Praxis
I grew up with parents who discovered my talent in painting, drawing, and sculpture during kindergarten and in my middle and high school years. I was admitted to a prominent art program affiliated with an ICT-oriented research institution. I further developed my design, aesthetic, and creative thinking skills, as well as my digital literacy, during my undergraduate studies. While pursuing master’s and Ph.D. degrees in the related fields of STEAM, and a full-time faculty appointment with a few ICT-intensive institutions for approximately a decade, I was able to experience transnational education, apprenticeship, and visiting scholarship with elite art institutes and research-oriented universities in Europe, North America, and Asia, including the United States, Germany, Italy, France, Spain, China, and South Korea (Bai & Nam, 2020, 2022, 2023, 2024).
In this way, I could articulate my knowledge and skills in art. More recently, I have taught APT, flash video, and diverse design-related courses, helping undergraduate and graduate students understand STEAM theory and praxis. One of my prominent recent roles in this area has been developing the STEAM identities of young college students and adult learners. While many students participated in elite art programs and developed professional identities as art designers and animation production technologists, students from low-income families often felt confused about their learning goals. They were excluded from the mainstream STEAM society. Text-to-video models may impact these existing structural problems (Bai & Nam, 2024; Gan & Bai, 2023).
Potential Human Capital Crisis in STEAM Academia and ICT Industry
Recently, an emerging topic in STEAM and ICT has been the potential of text-to-video models to address the human capital crisis across various academic and professional labor markets. Archival records reveal potential crises that have alarmed the STEAM society, particularly animation production technologists. For example, an archive record in Companies may leverage Sora to produce advertising videos that swiftly adapt to market changes and create customized marketing content. This not only reduces production costs but also enhances the appeal and effectiveness of advertisements. The ability of Sora to generate highly realistic video content from textual descriptions alone could revolutionize how brands engage with their audience, allowing for the creation of immersive and compelling videos that capture the essence of their products or services in unprecedented ways (Liu et al., 2024, p. 23).
Another archival record in The threat to creative jobs is a pressing concern, with fears that AI advancements could make human artistry non-competitive. A Reddit user voices distress about the future of creative education: “As a high school senior going into 3D modeling for college, I’m absolutely terrified that my education might be worthless in a matter of years.” This sentiment is echoed by another user, who fears that the creative industry might cease to exist, affecting not only artists, but other industries as well. (Mogavi et al., 2024, p. 4).
Additionally, an archival record in
Many advanced countries in Europe, North America, and the Asia-Pacific region, including China, South Korea, Japan, and Singapore, have been participating in cosmopolitan AI market competition, and STEAM programs have grown as a strategic tool. For example, in South Korea, STEAM education has been pivotal in fostering the next generation of human capital through art studio-based learning and school-community partnership programs, which have contributed to the development of elite art schools and institutes at all levels (Paek, 2020). Meanwhile, in Japan, there has long been gendered politics between men and women in academia. However, STEAM education has created a positive social movement, encouraging numerous girls and women to envision their career prospects as art designers or STEAM professionals (Kijima & Sun, 2021). Additionally, China has created positive sociocultural legacies through STEAM education. High-profile universities and polytechnics have played a crucial role in establishing substantial STEAM infrastructure, promoting a knowledge-based economy and human capital development. This, in turn, enables the nation to achieve a higher level of economic success through exceptional educational development. Many technical vocational education and training (TVET) students and adult learners expect to become art/interior/fashion/industrial designers, media engineers, social media app developers, and animation production technologists (Gan & Bai, 2023).
Despite this positive global educational movement, digital capitalism is deeply ingrained in the STEAM industry. Thus, students enrolled in STEAM programs may be marginalized within diverse professional industries, and very few elite art students enrolled in prestigious universities, polytechnics, and art institutes will succeed in the job market competition. While many lower-tier universities, TVET institutions, and adult learning and continuing education programs are concerned about their sustainable development and capabilities (Bai & Nam, 2024), the myth of the knowledge-based economy and human capital will no longer be part of the STEAM ideology ( Gan & Bai, 2023). Thus, governments and institutions must support general STEAM students in preserving their digital literacy and technical proficiency. From these perspectives, numerous students from STEAM majors must make extra efforts to cultivate self-efficacy, artistic knowledge, and aesthetic tastes beyond academic rigor to obtain college admissions from high-profile universities. In this stratified realm, it is challenging for high school students to develop a positive motivation for choosing STEAM-related majors (Bai et al., 2025).
Finally, in reflecting on my personal experience and close interaction with individuals in the STEAM industry, I witnessed this threat even before OpenAI’s announcement of Sora. Back in 2021, a female colleague of mine had her own interior design company, employing many graphic designers and IT software engineers who specialized in renovating houses and designing interiors. However, in 2023, her company closed because of the rapid evolution of AI. She became an employee at another interior design studio, one smaller than her own had been. However, she has recently begun to feel anxious about her employment status because of Sora. Thus, the job market crisis in the STEAM industry is already visible. Furthermore, academic programs related to STEAM face challenges in the economic structures of the contemporary capitalist and neoliberal educational market. My current institution’s academic leaders have recently discussed downsizing or integrating existing STEAM programs, with APT being one of the target programs. If this scenario happens, academic faculty members in the program will have to move to other programs or terminate their academic careers. Undoubtedly, they will also be able to seek positions at other institutions, but they may encounter similar challenges in the current AI era. This structural problem also impacts students’ learning goals and future career prospects. A recent conversation with a female graduate student is a prominent example of how a STEAM-related major can lead to anxiety about one’s future career. She is currently considering whether to continue in or drop out of her graduate program amid rapid changes in the political, economic, and sociocultural climate.
Relevant Issues for Brian as an Academic Editor and a Business Entrepreneur Trainer
I hold bachelor’s, master’s, and Ph.D. degrees in the related fields of BME and ELPS, with other credentials, such as minors and professional and graduate certificates in communication, English as a Second Language (ESL), global leadership, and cultural studies. I have been serving as an academic editor, editorial board member, and committee member for multiple journals, intellectual societies, and professional associations. I have also been training and consulting individuals in BME, ESL, and especially in the field of global business and educational entrepreneurship.
To the best of my knowledge, and in my role as an academic editor, I position myself as a pragmatist and gatekeeper for academic knowledge, feeling obligated to safeguard the academic property of researchers and respect their ethical research practices and moral commitments within the academy. From my perspective, I can somewhat distinguish between human-written texts and those generated by AI text-to-text models. I may have some biases, but AI chatbots often fail to use adverbs, relative pronouns, and conjunctions properly. In contrast, human writers actively employ them to convey their critical, creative, and innovative thinking. Of course, while undertaking my initial review to determine whether to proceed into the review phase, I often recognize minor grammatical mistakes, punctuation errors, and lapses in clarity, focus, and engagement in publishable scholarship. This issue even happens with academic writing work from native speakers. However, I value their logic, conservative efforts, and methodological considerations in the quality of their research reports.
By taking a business entrepreneur trainer role, I reflect on my involvement outside my academic boundaries. I have experience in helping develop strategic plans, recruiting and training employees, and serving as a consultant to business entrepreneurs in Europe and Asia. I cultivated a sense of business ethics that seeks to respect partners and treat them fairly, as I would like to be treated. Thus, my motto entails mutual trust, commitment, cooperation, and positive interpersonal and intercultural communication with business partners. These elements are crucial because my business ethics are closely intertwined with sales and marketing, making partners and customers of utmost value for my enterprise’s success–it is all about human-to-human interaction in social networks for mutual benefit. Making a profit is one of my long-term goals, but gaining partners’ trust is another. Thus, integrity, honesty, candor, and trustworthiness are crucial to business ethics. However, the emergence of ChatGPT could bring out different aspects of business ethics.
Initially, I shall reflect on the positional statements regarding the use of Generative AI models in the academic ecology. Since the 2022–23 academic cycle, the global academy has witnessed phenomenal growth in STEM and HE research seeking to understand the emergence of ChatGPT and its ethical implications for researchers and academic writers. As an experienced academic editor, I have observed that many natural scientists, STEM and/or STEAM scholars, and ICT researchers in fields such as mechanical engineering, biomedical technology, and astronomy have closely interacted with AI and heavily relied on its functions to conduct their research. They can be considered pioneers who have expanded scientific knowledge and facilitated human life in various ways. However, although Generative AI models are practical instruments in many ways, editorials have expressed growing concerns about ethical dilemmas, biased responses, and issues related to human intelligence (Nam et al., 2024b, 2025a, 2025b; Nam & Bai, 2023).
As a pragmatic business leader, I believe that business ethics are closely connected with pragmatic leadership and work ethic, shaping an organization’s success in many ways. In modern corporate and market economies, pragmatic leaders promote moral behavior, using their wisdom to help followers express their needs and shared interests. This way, organizational members develop a sense of belonging and work toward common goals. Therefore, a pragmatic approach to business ethics emphasizes how a leader’s actions can yield optimal results by mitigating conflicts and other social issues within an organization (Nam, 2025; Nam et al., 2019, 2022, 2025a).
Pragmatist Research and Business Ethics and Grounded Rules
A recent editorial in Nature Machine Intelligence (2022) discussed the early stages of the ChatGPT era and raised concerns about ethical issues in academic publishing. In this context, a persistent challenge for experimental AI tools is that they have difficulty understanding complex scientific knowledge and analyzing diverse cases and results. This editorial also argued that scientific research is facing a crisis; members of scientific societies may encounter ethical dilemmas that need to be addressed at AI conferences and through academic journals, which should encourage more experiments with ethical and responsible publication practices. Additionally, Nature Biomedical Engineering (2022) expressed similar concerns about scientific data and analysis in biomedical engineering. [C]ontextual learning is nearly all you need…Graph neural networks and transformers taking advantage of contextual information and large unannotated multimodal datasets are redefining what is possible in computational medicine…While biomedical engineers need to spend a significant amount of time interpreting data through traditional machine learning algorithms, they struggled with their tasks because of the “complex spatial relationships within medical images (p. 1319).
Editorials in other journals, such as Nature (2023), Nature Astronomy (2023), Nature Biotechnology (2023), and Nature Machine Intelligence (2023), also predicted that many scientists might overlook specific experimental procedures in scientific research, especially scientific writing and data management in the natural sciences and STEM fields, including astronomy, mechanical engineering, and biotechnology. AI chatbots could pose transparency issues due to a lack of clear ground rules for developing academic manuscripts, expressing uncertainties and moral dilemmas about whether they should use generative AI models as writing aids. Specifically, understanding intellectual property remains unclear. Notably, Nature Biotechnology (2023) provided an illuminating example of both positive impacts and potential challenges in academia. AI has been vital in biotechnology, yet the influence of AI chatbots on scientific research can bring many obstacles despite their valuable roles in the biomedical sector and traditional methods for protein engineering that enable interactive mutagenesis and sequencing. This editorial also expressed caution about how AI trains data to avoid potential biases related to disease variants, emphasizing that scientists will need to work with AI and minimize any biases the AI’s logic might generate during real-time interactions.
Additionally, an editorial in …authorship is an ethical issue of significant importance in scientific journals. A recent publication in Nature stated that an AI chatbot cannot take responsibility for the article’s claims. One of the most renowned AI-powered chatbots is ChatGPT…The primary objective of a scientific article is to convey new information that is supported by evidence, and all claims made by the authors should be thoroughly scrutinized before being published (Kim, 2023, pp. 1–2).
By the same token, another editorial in
A few qualitative experts in the given body of literature raised critical questions about the relatively limited scholarship that seeks to increase a comprehensive understanding of specific research design processes. In response, Chubb (2023) contemplated the use of AI for qualitative data analysis by demonstrating communication processes with chatbots. The previous study provided an overview of AI tools for qualitative analysis, tested how chatbots responded to data analysis, and demonstrated how to control false answers or illusions. Whilst the previous study offered some possibilities of using AI for qualitative research without vignettes, ethical complexities and limitations coexist in academic research, especially when qualitative researchers neglect to respect knowledge translation. Similarly, Hayes (2025) argued that LLMs could augment human intelligence and expertise through textual analysis. However, persistent ethical issues include “data security and bias,” in which qualitative research should “demand careful oversight and transparent reporting, emphasizing the importance of maintaining scholarly integrity” (Hayes, 2025, p. 1). Despite the growing attention to the use of Generative AI models, grounded rules are certainly significant for scholars in IQR, particularly given the praxis that should present creativity, novelty, and academic property in publishable scholarship (Nam et al., 2024b; Nam & Bai, 2023).
Despite academic journal editorials’ and HE magazines’ concerns about the challenges and opportunities of using ChatGPT, many business magazines have praised this platform as a practical instrument and have anticipated continual growth in various industries. For instance, Oktavia Catsaros from
Furthermore, Eric Mayhew, a columnist from
On the other hand, Wendell Wallach, a columnist for
Despite traditional models of business ethics and practical leadership philosophies, the rapid development of AI has raised increasing concerns about integrity in business ethics. In my view, well-known global business magazines have overlooked the human rights approach to AI and justice for all. As mentioned earlier, IT industry leaders and their companies in Silicon Valley (e.g., OpenAI, Google, Microsoft, and Deepseek) have focused exclusively on enhancing AI infrastructure and upgrading their software. They present their products as practical tools that contribute to humanity’s future while also boosting profits and financial success. People from diverse intercultural and socioeconomic backgrounds may view these practices differently, depending on their familiarity with global business and entrepreneurship within their respective societies and countries (Kulich et al., 2020). Notably, rising concerns include misinformation, data interpretation issues, and sanctions. In this context, I have highlighted the limited roles of public intellectuals and business leaders, as well as the constraints they face in criticizing the neglect of ethics and morality in the age of AI. Therefore, we need to pay closer attention to existing corporate and government oversight systems and work towards developing more explicit rules, policies, and practices.
Concluding Remarks: Methodological Implications for IQR on Generative AI Models
Qualitative researchers select their methodologies and methods based on their subjectivities, which entail reflexivity, positionalities, social representations, and cultural practices within the context of naturalistic inquiry (Creswell, 2013; Denzin & Lincoln, 2011; Lincoln & Guba, 1985). From that standpoint, qualitative experts called upon scholars to muse on “starting out on their journey in applying realist research approaches” (Wong, 2015, p. 1) and “the prioritization of social impact” on posing “a series of methodological challenges,” in which “having provided important advancements in scientific knowledge and evidence, are not enough” (Soler-Gallart & Flecha, 2022, p. 1). In other words, scholars are aware of “rigorous and ethical research practices” and commit their works “through detailed methodological descriptions” about their empirical results or conceptual findings (Poth, 2019, pp. 1–2). However, reporting methodological insights and implications has so far been limited, particularly in terms of ethical standards and trustworthy, transparent, and reproducible practices, despite the long-term “pioneering qualitative spirit” and “inclusive space for sharing and learning about qualitative methods” (Sousa et al., 2018, p. 1). Therefore, as an exemplar, we aimed to design a meta-autoethnography of metadiscourse on the use of Generative AI models. We attempted to demonstrate specific research procedures and address key methodological issues relevant to and applicable in bolstering qualitative-based naturalist inquiry.
Reflecting on our subjectivities and interactions with community members, a female colleague of the first author owned an interior design company that employed numerous graphic designers and ICT software engineers. However, in 2023, the colleague’s company closed due to the rapid evolution of AI, particularly in text-to-video models, which could contribute to growing concerns about technological unemployment (Nam et al., 2025a; Peters et al., 2024). There has been a growing quantitative body of literature on the impact of Generative AI models on STEAM and HERD. This surge has potential implications for human intelligence in the current academic and business job markets (García-Peñalvo, 2023; Kim, 2023; Kumar et al., 2024; Rayner, 2023; Shen et al., 2023). Accordingly, we urge future qualitative researchers to devote more attention to these emerging structural issues and engage with the public to develop effective coping mechanisms for humanity through their IQR and STEAM praxis.
We presented the fundamental nature of qualitative research and manuscript development in publishable scholarship at a general level. We then addressed issues involving IQR, more specifically, thereby helping readers gain insight into our intentions for designing a meta-autoethnography of metadiscourse on the use of Generative AI. In turn, we expanded scholarly conversations about the role of subjectivities in qualitative research and disputed the challenges and opportunities for enhancing qualitative studies. Accordingly, we revealed our subjectivities to acknowledge our biases toward different disciplines as co-joint authors. Persistent concerns of academic editors, reviewers, and authors, who deal with qualitative methodologies and methods vary, including the originality and novelty of the manuscript, theoretical relevance, methodological applicability, research ethics, trustworthiness, and practical implications, among other essential elements in publishable scholarship (Creswell, 2013; Egbert & Sanden, 2015; Merriam & Tisdell, 2016).
Nonetheless, we attempted to solve these critical issues by using liberal, practical, and co-constructive approaches to promote scholarly conversations about pragmatist ethics and critical pedagogy; we considered the methodological rigor through philosophical assumptions and interpretive frameworks (Creswell, 2013). We focused on promoting scholarly conversations about pragmatist ethics and critical pedagogy as the ontological and epistemological approaches for conducting this research. In turn, we attempted to gain a comprehensive understanding of the methodological relevance and applicability by designing a meta-autoethnography of metadiscourse. More specifically, we considered how to approach hermeneutic data collection and analysis strategies. We discussed key concerns regarding the contextualization and trustworthiness of the research.
Finally, some notable limitations have remained. Although STEM also encompasses engineering and mathematics, we acknowledge that our expertise and reflexive knowledge extend beyond these areas. Currently, growing concerns include scientific knowledge and AI competition between the U.S. and China, which are developing geopolitical tensions and crises in academic diplomacy (Altbach & de Wit, 2023; Nam et al., 2025a, 2025c, 2025d). Since our scope solely covered critical pedagogy, business and research ethics, and methodological rationale, and its subsequent implications for IQR on generative AI models, we seldom discussed the geopolitical impacts on AI and its brain circulations. Therefore, future scholars should consider the limitations of this paper and continue to advance social scientific knowledge.
To conclude, we demonstrated the methodological implications by triangulating our own meta-autoethnographic texts and the chosen hermeneutic data with the potential metadiscourse themes, alongside our philosophical assumptions and interpretive frameworks. Therefore, we shed new light on IQR on generative AI models, providing concluding remarks for a broader audience. Overall, we offered our reflexive insights on the most pressing issues and structural problems in the current era of generative AI.
Footnotes
ORCID iDs
