A Philosophical Inquiry Into Utilizing ChatGPT Through an I-Thou Framework

XIAO DONG
Western University

BETTY ANNE YOUNKER
Western University

June 2025

Published in Action, Criticism, and Theory for Music Education 24 (3): 109–37 [pdf]. https://doi.org/10.22176/act24.3.109

Abstract: Research into using AI for editing doctoral dissertation work in music education and a subsequent review of literature prompted this collaborative investigation. Specifically, this paper examines ChatGPT-Human collaboration in doctoral dissertation writing and editing through the lens of Martin Buber’s (1958) I-Thou relation. Constructed through dialogical discourse (Bakhtin 1981), our (the supervisor and the supervisee) voices interact with the intent to explore: (1) the ways ChatGPT was utilized for editing the supervisee’s dissertation and how reflexivity influenced the process, (2) the impact that ChatGPT-Human collaboration has on the supervisor-supervisee role shift, (3) the ethical considerations, including the supervisee’s voice, authorship, and agency that can be impacted in response to such shifts, and (4) whether such shifts and impacts contribute to specific aspects of our pedagogical values as teachers in the field of music education. This paper offers insights into the practical application of AI in music education and advocates for further honest dialogues regarding the utilization of ChatGPT.

 Keywords: ChatGPT, I-Thou relation, dissertation writing, authorship, voice

Within the evolution of AI, large language models (LLM) have been developed including Chat Generative Pre-Trained Transformer (ChatGPT) (Dergaa et al. 2023). With these developments, discussions have emerged in the academy about the possible benefits, threats, ethical considerations, and impact on the authenticity and credibility of academic work. ChatGPT has been regarded as a tool to assist in the writing of research papers, including doctoral dissertations. Such use could shift the roles and tasks of the supervisor and supervisee. To deepen our understanding of supervisor-supervisee relationship dynamics in the context of ChatGPT usage, we selected Martin Buber’s I-Thou relationship as our theoretical framework. Unlike the traditional teacher-student model, in which teachers view students as receivers of knowledge, the I-Thou framework emphasizes a more fluid and dynamic relationship in which the roles of learning and guiding are fluid and dynamic. In this model, dialogical interaction plays a central role, creating space for relationships and understanding of self and others to grow. Through the lens of I-Thou relations, several questions guided the inquiry: What might be the changes in roles of the supervisor and supervisee? What might be the impact on the working relationships between the supervisee and supervisor? What might be the impact on the supervisee’s voice, authorship, and agency? What pedagogical values might be questioned or affirmed as the above shifts and impacts are felt?

Focus and Guiding Questions

With the above curiosities, the following questions guided our reflective thinking throughout the writing process:

  1. In what ways was ChatGPT utilized for editing the supervisee’s dissertation, and how did reflexivity influence the editing process?
  2. In what ways, if at all, do roles of a supervisee and supervisor shift during the introduction of ChatGPT?
  3. How do the shifts impact, if at all, the professional relationships between supervisee and supervisor?
  4. In what ways, if at all, does utilizing ChatGPT as a tool impact the development of a supervisee’s voice, authorship, and agency?
  5. In what ways, if at all, is one’s pedagogical values impacted as a result of discussions about and reflections on the above?

With a focus and guiding questions, we (the supervisee and supervisor) engaged in a dialogical discourse that included learning from and informing the other. During these conversations, the roles of supervisor and supervisee shifted as the supervisee guided the supervisor about how to “teach” ChatGPT and provided suggestions from an “ideas” perspective. The supervisor provided suggestions from an “ideas” and editing perspective. The discourse was iterative as we began discussing the possible use of ChatGPT in the writing and editing processes of a doctoral dissertation. As the dialogue continued, two aspects of our interactions have deepened and broadened beyond the practical utilization of ChatGPT. First the discussions have engaged us reflexively as topics related to roles, relationships, ethical considerations, and possible impact on one’s pedagogy have emerged. Second, as we began reflecting and writing, and reflecting on each other’s writing we experienced a probing into each other’s thinking and subsequently a curiosity to unpeel each other’s meaning—meaning as constructed throughout the experience. To that end, the discourse grew dialogically, and our relation began to strengthen as I-Thou.

Review of the Literature AI and ChatGPT in the Field of Education

Recent progress and expansion on generative AI (GAI) models have been considered a “game changer” as they are built on deep learning techniques, trained on large datasets, and able to generate text, images, music and video (Maphoto et al. 2024). One of the most powerful GAIs is ChatGPT from OpenAI. In this section, we review the current literature about the use of AI and ChatGPT in education, synthesizing how teachers, students and researchers perceive the benefits and challenges of AI technologies.

Benefits of AI. UNESCO published a report in 2021, titled AI and Education: Guidance for Policy Makers, in which the leverages and risks of AI are thoroughly discussed (UNESCO 2021). From the teachers’ perspective, AI tools as assistants can complete several kinds of tasks, including answering students’ questions; providing feedback to students’ assignments; and designing course syllabi, activities, and rubrics of assignments. This could shift energy and time from those kinds of tasks to ones that involve acquiring more complex understanding about innovative pedagogical perspectives, and knowledge and skills, all part of professional development (Grassini 2023; Maphoto et al. 2024). The use of AI tools could allow students more time and energy for tasks that require solving problems, doing assignments, and brainstorming for creative activities. Additionally, students could receive supplementary support from AI machines, e.g., using the tutor program to instruct students’ basic subject material (Fahimirad and Kotamjani 2018). With the advancement of programs, the level of support and sophistication has increased. One example is ChatGPT. Its features go beyond answering basic questions to providing information at a greater breadth and depth. This allows for learners to be introduced to new topics and to interpret, analysis and synthesize new information thus, to engage critically with the material.

Since ChatGPT is a dialogue-based AI, learners can, with appropriate and specific prompts, personalize their learning processes (Baidoo-Anu and Ansah 2023). Personalized assistance allows for teachers and students to explore their educational practices in different ways, including their roles as teachers and learners and the level of engagement as experienced. In addition to this identified assistance, AI has altered the way the public in general communicate and engage with each other moving from sharing of knowledge in print form to dialoguing globally through technology.

Challenges of AI. With benefits also come challenges. Much has been written about the inaccuracies and biased information generated by AI. Several authors have reported “AI hallucination,” where ChatGPT generates plausible and convincing yet incorrect answers, leading to the spread of misinformation (Mondal and Mondal 2023; Semrl et al. 2023). Dergaa et al. (2023) emphasized that the integration of false and biased information led to “unintentional plagiarism and/or the misattribution of concepts” (617). Moreover, when Semrl et al. (2023) examined the feasibility of using ChatGPT to assist research, it turned out that ChatGPT lacked the ability to provide citations to support the arguments. Finally, it has been noted that ChatGPT might create references that do not exist. We, as educators and researchers, need to be mindful, therefore when utilizing such programs. Bowman (2022) likened the interaction as chatting with a wise intern whose perception of truth is clouded by the desire to please you.

In response to the challenges, researchers have queried about students’ level of engagement when using AI as learning tools, and the possible growth as independent learners. To address these queries, researchers have suggested that teachers discuss with their students learning strategies that are effective and require higher order thinking skills. In addition, it has been suggested that teachers rethink the design of assessments and consider creating ones that require increased use of critical and creative thinking skills and spoken expression (Dergaa et al. 2023; Maphoto et al. 2024).

Based on the synthesis of current literature, there is a deficiency of practice in exploring AI technologies in education. Thus, researchers are encouraged to investigate practical cases in diverse educational contexts and contribute to a deeper understanding of the use of AI. There is caution, however, that though AI and ChatGPT can be useful tools in education, precautions are necessary to ensure the efficiency and effectiveness of AI technologies. As Mondal and Mondal (2023) suggest, “it is important to maintain human judgement and critical thinking” (3605) in the use of AI.

AI and ChatGPT in Academic Writing

In terms of using generative AI in academia, scholars interact with AI to assist with multiple phases of research, such as framing research questions, searching for related literature, choosing appropriate methodologies, analyzing data, editing, and proofreading texts (Dergaa et al. 2023; Mondal and Mondal 2023; Semrl et al. 2023; Wise et al. 2024). However, ChatGPT’s performance on these various phases is not always satisfactory. In addition to generating inaccurate content as aforementioned, researchers noticed that the process of communicating with ChatGPT is inconsistent. Specifically, when researchers need ChatGPT to continuously generate appropriate content in response to a series of prompts, they need to reactivate or rethink the prompts to ensure that ChatGPT does not generate less pertinent information. Indeed, the quality of ChatGPT depends on the user consistently refining and re-tooling the instructions and prompts. Researchers who explored how different doctoral students used generative AI in their writings found that “the iterative coordination of writing processes with AI assistance typically yields better results than linear methods” (Nguyen et al. 2024, 13). This is to say that individuals should coordinate their writing processes, shifting fluidly from sequential approaches to more concurrent strategies, including the refining of instructions and prompts in response to what was generated. As well, as students become more familiar with AI strategies, their ability to take advantage of the benefits will increase (Maphoto et al. 2024).

Even if researchers are well-equipped with technology-related knowledge and strategies, ethical considerations including voice, authorship, agency, and authority should be interrogated as they are critical for the value and authenticity of the research (Amirjalili, Neysani, and Nikbakht 2024). In terms of authors’ voices in academic writing, Amirjalili, Neysani, and Nikbakht (2024) adopted mixed qualitative and quantitative methodologies through comparing human writing and AI-generated text to explore how authors construct their voices in writing. Specifically, Amirjalili, Neysani, and Nikbakht (2024) drew from Helms-Park and Stapleton’ (2003) empirical study, utilizing a “Voice Intensity Rating Scale” to measure multiple elements of voice, such as “content, structure, language use, vocabulary, mechanics, self-identification, assertiveness, repetition of the main idea, and authorial presence” (3). In addition, the Text Inspector tool was used to statistically measure the ratio of different words to the total number of words, examining the linguistic diversity in text. Findings indicated that human writing exhibited a more diverse and richer vocabulary, nuanced voices, and effective rhetorical use. Therefore, human writing demonstrated a strong sense of authorship as it was constructed in a more subtle, unique, and personalized way.

Many researchers see the possibilities of utilizing AI in academia but with much needed discussion across a variety of topics including authorship and acknowledgment of utilizing AI. One publisher, Springer Nature, advised not to credit ChatGPT as the author as it cannot take responsibility for the content it generated; this is a human attribute and responsibility (Dergaa et al. 2023). Researchers are encouraged to document their use of AI by acknowledging sections in their papers that have been impacted. In the meantime, there is a call for creating policies that guide the use of AI in the academy.

Teacher-Student-AI Relationship

Teacher-student relationships in the era of technology has been transformed. In a traditional model, teachers as experts play dominant roles in the learning process as they possess knowledge and skills in the field. However, with the development of learning assisted tools such as ChatGPT, teachers are no longer the sole bearers of knowledge. This concerns some educators as they see their role as being the experts possibly becoming extinct. Grassini (2023) argued that instead of just being bearers of knowledge, teachers as human beings can understand students’ emotional status, thus provide more targeted emotional support such as empathy and encouragement. This emotional interaction is significant for students’ motivation and learning outcomes, and is a skill set that AI is currently lacking. The possible shift then in roles of teachers and students begs an examination as it may impact pedagogical practice.

Guilherme (2019) interpreted the teacher-student relationship in the era of AI through the lens of Martin Buber’s I-It and I-Thou relations. According to Guilherme (2019), the I-Thou relation “has been described as a dialogue and an inclusive reality between individuals, a reality in which one makes space for the Other to be who [they are]” (49). This is to say that the irreplaceable value of teacher-student relationship is the dialogical interaction between the two participants. In doing so, students not only absorb knowledge and skills, more importantly, they begin to strengthen voice, authorship, and agency, thus potentially begin to engage in an I-Thou relation. This human interaction includes spaces for critiquing, exchanging intriguing thoughts, and reflecting throughout the process.

In human-AI interactions, an I-It engagement occurs as learners play the dominant role through creating and supplying prompts as more distinct and conversations are desired. AI is viewed objectively as an assisted tool—one that provides responses to human-generated prompts. This differs from two human beings engaging and interacting, responding and debating (engagement through I-Thou relation) to more of a traditional model of teaching, in which the responsibility of the teacher is to pass on knowledge to the student and guide them in their learning. With AI, one could view the learner as being the knowledge bearer who guides AI with refined prompts to elicit relevant and meaningful responses.

Summary

We have presented a review of literature that includes the benefits and challenges of AI in educational settings and academic writing, noting that while benefits do exist, further research and discussion is needed. Noting the challenges, it is recommended that policies are generated to serve as guides for when AI is being utilized. Publishers have begun to recommend that AI is not listed as an author but rather a tool that assists in specific ways, and to acknowledge sections of the paper that have been impacted. Emerging from the benefits include an increase of space in which teachers (including supervisors) and students (including supervisees) could focus on more complex tasks that require enhanced critical thinking and could experience shifts in each of their roles, which might strengthen an I-Thou relation.  With this strengthening one’s voice, authorship, and agency could be identified and strengthened as well.

Definitions

For the purposes of this paper, the following definitions for voice, authorship, and agency are utilized. Voice is identified as an expression of the author’s understanding and critical thinking, and a distinctiveness that encompasses their “attitude and intellectual journey” (Amirjalili, Neysani, and Nikbakht 2024, 2).

Authorship is defined as encompassing “ownership, accountability, and the integrity of ideas” (Amirjalili, Neysani, and Nikbakht 2024, 2), which can come with maintaining one’s voice and understanding the responsibility of putting forth and espousing ideas. It has been noted that ChatGPT cannot be responsible for the research and thus cannot replace the actual complete work of the researcher (Dergaa et al. 2023). ChatGPT can certainly serve as an aid, a tool, as noted above, however it is critical that transparency about how the tool is used and for what needs to be clear in the delineation between the researcher’s creative and critical work and the “work” of ChatGPT.

Agency is one’s capacity to make choices in a mindful way, to think reflexively, and to act independently while being cognizant of possible structures that may impede their capacity to act. Such structures could include social-economic factors, gender, religion, ethnicity, or cultural constraints. An integral part of the human condition is to act, choose, and imagine, to construct meaning through agency (Brockmeier 2009). Brockmeier draws on the work of Bruner (1986) when examining the construct of humans creating and understanding meaning across cultural contexts and recognizing this as part of the human condition. Brockmeier (2009) offers an understanding of agency through the writings of Bruner (1990) and Holzkamp (1983) as “the agentive dimension of human subjectivity” (218). Humans have the capability to influence, choose, and possibly change their current conditions as experienced. With respect to this paper, shifts in roles and relationships, and pedagogical practices could contribute to strengthening a supervisee’s agency and to a supervisor’s value of such agency. This would reflect relationships that are the anthesis of a master-apprentice or other like traditional relationships that are unidirectional and in which the only action taken is based on choices determined by the teacher. A relationship in which voice and agency are valued involve identifying choices for reflection, implementation, and evaluation.

Philosophical Framework

For the purposes of this paper, we examined ChatGPT-Human collaboration in doctoral dissertation writing and editing through the philosophical lens of Martin Buber’s (1958) I-Thou relation. This philosophical framework provided guidance as we examined the focus of the paper and crafted the questions that guided our discussions, reflections, and writing.

Buber I-Thou

Buber (1958), through his theoretical framework of I-Thou, fundamentally embodies relation and the inner nature of that relation, one that is beyond a notion of a sensation or to call it as, or identify it with, a feeling. It is the immediacy of relation. I-Thou encounter, according to Hibbard (2017), “is subjective in nature and characterized by reciprocity; two beings enter into a relationship as an entity, not the sum of its individual qualities” (38). Therefore, the inextricable nature of I-Thou is that of being a primary word, reflecting on how individuals relate to each other as human beings—of being human, of the betweenness with other, of establishing a world of relation, and of each impacting, all without judgment of qualification. Buber writes not of describing or experiencing but of “bodying forth” during which one discloses. This differs from I-It, in which there is no relation, no betweenness with other, and no concern for nor impact on It. Thus, within the I-It experience, the distance always exists between the subject and the object.

In I-Thou relation, relationships are built and the being with others strengthens. As individuals are with others without judgment or qualification (Buber 1958), they can come to know and be with each other through uncovering the other while reflexively uncovering our biases, ideas, and assumptions (Benedict 2021). It is through this interaction—the moment of the interaction—that guides the participants, and it is during the interaction that thinking and knowing reflexively occurs. As interaction and participation in the dynamic, living process occurs—an I-Thou relation is experienced. Benedict (2021) continues with this line of thinking as she writes about how the essence of relationships with others is what matters. Buber (1958) uses the example of a tree to illustrate the distinction between the I-It and I-Thou relationships. The key lies in how a person perceives the tree. When someone views the tree as an image, describes its movement, or studies it scientifically, the tree is treated as an object, placing the person and the tree in an I-It relationship. However, if the person instead focuses on their relationship to the tree, “rather than any inherent quality in the tree itself” (Hess 2021, 74), the relationship shifts from I-It to I-Thou.

This relation requires a pedagogical shift from teaching content to students to interacting reflexively. It is a way of being and interacting, working with students intentionally (Smith 2012). This involves building relationships and trust, being aware of other’s needs and wellbeing, and entering the betweenness with practices of relational care (Noddings 2005). For example, in working with traumatized students, Hibbard (2017) reflected on her self-identity as a white teacher in an African American community in Detroit, acknowledging her own internalized racism and the deficit assumptions she held about her students. As Hibbard (2017) stated, “the more honestly I met myself, the more I was able to meet the students in the same way, and they began to reciprocate” (44). Trust does not emerge spontaneously; it requires effort to build. Teachers need to be genuinely present in the teacher-student relationship, authentically engaging with themselves and inviting space for students to join in a mutual relationship (Hibbard 2017, 2021). This engagement impacts I-Thou relationships as participants interact while learning through constructing, inquiring, expressing and communicating (Dewey 1938).

Guilherme’s (2019) writes about the potential shift in relationships between educator (supervisor) and student (supervisee) because of the increased role of and attention on technology in the learning process. They contend that the shift is from an I-Thou to an I-It relationship. This shift impacts the relationship, now viewed as becoming detached, and that the connectedness between the teacher and student “has decreased or become impaired” (47), thus the capacity to bond has diminished.

Conversely, we, as authors of this paper, argue that the shifts as experienced are within the roles played during the process, thus impacting and indeed, strengthening the relationship between supervisor and supervisee. As we reflected on those shifts, we became aware of the necessity to consider specific ethical concerns, including voice, authorship, and agency, as well as our own pedagogy and the fundamental value that informs our practice. Specifically, per Buber’s theory of I-Thou, we examined how utilizing technology such as ChatGPT during the writing process of a doctoral dissertation could shift the roles of the supervisee and supervisor and subsequently impact their relationships in terms of strengthening I-Thou. We argue that reflexivity (Benedict 2021) and relational care (Noddings 2005) as integral to I-Thou contribute to the quality of the relationship and thus impacts one’s pedagogical value and subsequently practice. This continues to be an exploration of “bodying forth” (Buber 1958).

Method of Dialogical Discourse

While data were retained via notes taken during our discussions our actual words were not retained verbatim as raw data. As we meta-cognized during our discussions it became apparent that our engagement during our exploratory journey was dialogical. What follows is a description of dialogical discourse followed by our reflections and discussion.

Dialogical Discourse and Reflexivity

Dialogic discourse involves voices interacting with the intent to explore the meaning of something. During the process, meaning is explored and clarified through collective negotiating, constructing, and discovering. This exploration and clarification occur through a dialogic access into our words which reveals semantic aspects of the word. As our understanding of the word is strengthened so is the dialogic relationship to the word (Bakhtin 1981). As supervisee and supervisor, we strive to strengthen but acknowledge the challenges that we need to recognize and interrogate. One is that perceptions of meaning and reflection on the perceptions can differ based on individual experiences. Second, each of our roles have traditional perceptions and that shifting those roles requires us to be mindful of why and how the shifts occur. Third, the supervisee identified English as her second language, while the supervisor identified English as her primary language. To strive for clarification and deeper understanding, reflection on each other’s words to gain access to the meaning of what is being offered and engaging in an iterative discussion was necessary.  While what we offer are reflections about the dialogical discourse and process and perceptions of that process, we acknowledge that reaching a level of being sensitive to the “essential attribute” of how each of our worlds are “seen and felt, ways that are organically part and parcel with the language that expresses them” (Bakhtin 1981, 367) is an ongoing journey as the supervisee continues with her dissertation writing. The ongoing journey essentially includes moving toward our languages as being an image of sensing and seeing the world and of being “one of many ways to hypothesize meaning” (Bakhtin 1981, 370). Finally, we acknowledge what Bakhtin (1981) refers to as “heterogloissa,” that which governs the meaning in any utterance. The situatedness of a word uttered defines the meaning thus a meaning could differ in a different context. We acknowledge, then, that the meaning of this current dialogical discourse is situated in the context as described.

During our interactions, we strove to find meaning of our experiences and pedagogical values as roles shifted, relationships altered, and agency was strengthened. The process as experienced was interactive and iterative, during the dialogue in meetings and between the meetings and writing. Each impacted the other, as reflected in the continued dialogue and writing. Camlin (2015) emphasized that real dialogue occurs when teachers and learners bring their unique perspectives into a “dialogic space” to explore diverse possibilities, rather than converging on a predetermined outcome. In our collaborative process, we aimed to think about and reflect on our practices with openness to flow. Similarly, Mitchell and Benedict (2020), engaging in genuine dialogue as a music therapist and a music educator, embraced the potential for surprise as they represented, interrogated, and interpreted their experiences both individually and with each other.

Through this process, as discussed below, the supervisee experienced an openness and strengthening of voice and thus agency. The supervisor experienced being a student as the supervisee instructed her in how to refine “teaching” ChatGPT. This aligned with what Freire (1970) addresses about the teacher-student relationship in dialogue: “the teacher is no longer merely the-one-who-teaches, but one who is himself taught in dialogue with students, who in turn while being taught also teach. They are jointly responsible for a process in which all grow” (61). The knowledge production was a dynamic and social process during which each of us offered multiple perspectives that were integrated into the dialogue and writing.

As we reflected on how we communicated with each other, we engaged reflexively examining the consequences of what and how we communicated, and why. How did our communication in action and what we communicated impact the other? What possible implications and biases became transparent? What did we uncover about self? What emerged in our dialogue was the realization that each was affected mutually and continually in our discourses. Each was accountable as we reflected on our roles and actions (Alvesson and Kaj 2017). Benedict (2021) builds on the role of listening and responding reflexively. She reminds us that as we talk with and make sense to each other and selves, we experience meaningfulness. During this process we ask ourselves, (1) “Who am I?” (2) “How have I become to be?” (3) “How do my actions with others encourage or prohibit engagements of inclusivity?” (10).

Description of the Evolution of the Paper and Reflections

Supervisee. During the development of my dissertation, my supervisor and I explored the use of ChatGPT to edit and enhance our reflections on the writing process. This collaboration not only improved the clarity and conciseness of my writing but also prompted the exploration of significant music education topics, leading to enriched discussions. Consequently, this paper retrospectively examines our collaborative editing process, the roles we each played in my dissertation writing, and the impact of these interactions on my development as a scholar. Specifically, it addressed how I established my academic ethos and navigated my authority, facilitated by our open-mindedness and critical examination of AI in music education. Eventually, our philosophies as educators were continuously constructed and re-constructed.

Supervisor. As I reflected on the supervisee’s discourse, it required me to think about and affirm the shifting roles as learner, teacher, and guide. These shifts impact relationships as they evolve, and it is in that space that one’s pedagogical values begin to take shape and be examined. For me, reflexivity and relational care (Noddings 2005) continue to be of value and thus shaped. This process required me to reflect on how reflexivity and care might be nurtured to strengthen relationships and increase affordances of voice, authorship, and agency in response to utilizing ChatGPT for academic writing.

Discussion

In the following section we address the findings within the context of each research question and include discussion and implications.

Question #1: In what Ways was ChatGPT Utilized for Editing the Supervisee’s Dissertation, and how did Reflexivity Influence the Editing Process?

Supervisee: The procedure and rationales of ChatGPT-human collaboration. In early February 2024, I met with my supervisor to discuss the data analysis section of my Ph.D. dissertation. It was the first time that I introduced ChatGPT to my supervisor as we were discussing the most effective and beneficial way to support my writing. Based on the experience working with my former supervisor, I noted that verbatim editing and proofreading were particularly labor-intensive for her due to English being my second language. Thus, I need much more grammatical correction and linguistic suggestions than merely feedback on writing content and structure. I tried to hire a professional editor to reduce my supervisor’s workload, allowing us more time to focus on critically discussing the ideas and improving the flow of the writing, and its depth and quality. However, the cost of professional editing was a significant consideration. Since the ChatGPT-4 launch in March 2023, I recognized new possibilities for writing assistance. Before the meeting, I had experimented with ChatGPT-4 and was impressed by its performance on many tasks, including editing text. I suggested to use ChatGPT to address grammatical errors, inappropriate vocabulary use, and issues of clarity and conciseness. Although my supervisor was initially suspicious of the AI assistance, she was open-minded and willing to experiment. This marked the beginning of our ChatGPT-human collaboration.

Following the meeting, we established a collaborative procedure with ChatGPT-4 for my writing, structured as follows:

  1. I independently complete the draft of the data analysis section.
  2. I then utilize ChatGPT-4 for the first-round editing and proofreading.
  3. My supervisor reviews the AI-edited text, providing insight and critical feedback
  4. Both my supervisor and I reflect on the AI-assisted process and discuss strategies to enhance the effectiveness of our ChatGPT-human collaboration.

These procedures are rooted in concerns for research integrity, authenticity, the preservation of the author’s voice, and fostering independent learning. My responsibility as a researcher includes ensuring the originality of the content and the authenticity of both the raw data and its interpretation. After completing my initial draft, I prompted ChatGPT to edit part of my writing. During the editing process, I actively managed the interaction with ChatGPT, leveraging both cognitive and linguistic resources iteratively. Subsequently, I clarified the specific prompts that I used during the process and revised them seeking better results. Though I kept reflecting on the AI-editing process, my language barrier as a second language writer limited my judgment of ChatGPT’s work; that is when my supervisor joined me to critically scrutinize ChatGPT’s editing from the perspective of a native English speaker. In our final discussions, we evaluated the structure and content of my writing, assessed any potential impact of the AI on my authorial voice, reviewed my supervisor’s comments, and strategized on improving ChatGPT’s utility. The ChatGPT-human collaboration process was inherently iterative, allowing us to revisit any step as needed to refine our approach.

Thick description of the ChatGPT-human collaboration process

Given that ChatGPT can process a document, I initially uploaded the draft of the data analysis for ChatGPT to proofread. However, since ChatGPT cannot directly modify the document, it instead summarized the key points and suggested improvement. Consequently, I decided to upload text section by section, requesting ChatGPT to directly edit the text.

My initial prompts concentrated on grammar, punctuation, style consistency, APA formatting, and overall clarity and coherence, aligning with the standards expected for doctoral-academic writing in music education. After submitting several paragraphs for proofreading, I observed that ChatGPT had reordered my statements, which disrupted the intended flow of thought and thus compromised my authorial voice. In response to ChatGPT’s performance, I adjusted my prompts:

  • You: “OK, for the following text I will send to you, please do not rearrange the order of my statement. Keep focus on clearing up grammarly [grammatical] issues, punctuation, improving clarity and coherence at the level of doctoral dissertation, but keep my own tone. Do you understand?”
  • ChatGPT: “Understood. I will focus on addressing grammar, punctuation, improving clarity and coherence, while preserving your original tone and not rearranging the order of your statements. Please go ahead and share the text you’d like me to proofread.”

ChatGPT adhered to these instructions initially, summarizing its process after editing:

  • ChatGPT: “This revision focuses on clarifying grammatical issues, punctuation, and enhancing coherence and clarity, adhering to the level expected of a doctoral dissertation, while preserving your original tone and the structure of your statements.”

However, as noted in the reviewed literature, ChatGPT sometimes deviates from the original prompts and thus generates content that is unrelated. This issue emerged several times during our collaboration. For example, when I submitted a paragraph for proofreading, ChatGPT instead evaluated the text. Once I noticed this deviation, I promptly reminded it to adhere to the original instructions, ensuring the focus remained on proofreading rather than text evaluation:

  • You: “No, that was not I meant, I needed you to continue to edit the text I sent with the same requirements we have discussed”.

Generally, ChatGPT was able to adjust its process and generate more satisfactory results. However, when ChatGPT was confused about the instructions, it asked for further clarification. For example, when I continued sending my text, the content may have confused ChatGPT. It questioned:

  • ChatGPT: “I understand your request but it seems there’s been a misunderstanding in how I can assist you further. If you have more text related to your dissertation or any other queries on which you need assistance, feel free to share!”

This response from ChatGPT indicated that generative AI has moved beyond where AI was merely waiting for instructions. Instead, it is capable of asking questions when it cannot understand the instructions or needs further clarification.

Based on my experience collaborating with ChatGPT-4, especially on my Ph.D. dissertation proofreading, I argue that ChatGPT performed productively. It adhered to the constructed prompts and generated academically rigorous text. Although there were instances of ChatGPT deviating from my prompts, it did not require much time or effort to bring it back on course. The other finding about ChatGPT’s proofreading abilities is about its linguistic patterns, such as fixed vocabulary selection and sentence structures. This is in line with the findings of Amirjalili, Neysani, and Nikbakht (2024), who noted that human authors had more diverse vocabulary and nuanced voice in their writing compared to ChatGPT.

When I shared my draft of data analysis section and ChatGPT-edited document to my supervisor, her feedback and comments laid the groundwork for our further reflection on improving my writing and collaborating with generative AI. Generally, my supervisor’s edits fell into two categories:

  • Comments Aimed at Content and Ideas:
    • Example:
      • Supervisor: “Does this reflect an interactive teaching approach? If so, perhaps another sentence that reflects interactivity that segues into the next sentence?”

This comment allowed me to reflect on my argument about the interactive teaching approach. If I insisted that an interactive teaching approach was embodied in the data, I needed to solidify the argument by adding another sentence to improve flow and clarity, as my supervisor suggested.

  • Comments Aimed at ChatGPT’s Proofreading:
    • Examples:
      • Supervisor: “Would you use the word ‘mechanics’? If so, leave it; if not, perhaps another word?”
      • Supervisor: “Perhaps ‘reflecting’? Not sure if ‘showcasing’ would be a word typically used by you?”

In our subsequent meeting, we discussed the feedback and comments in detail. My supervisor indicated she would review the use of vocabulary to guide reflection about the preservation of my authentic voice rather than being replaced by ChatGPT’s editing. She also emphasized the linguistic pattern in ChatGPT’s proofreading, noting that a set of vocabulary was repeatedly used, such as “showcase,” “underscore,” and “multifaceted.” Some of these words were not appropriate in the context of my research. For example, rather than revealing a deeper understanding or metacognitive development, “showcase” implies a surface-level demonstration. Additionally, certain words, like “albeit,” are no longer widely used in current academia. My supervisor’s insight into academic writing in the field of music education provided a valuable perspective to protect my voice and critically examine ChatGPT’s work. Her feedback highlighted the importance of careful vocabulary selection to maintain the integrity and authenticity of my academic writing.

In response to improving ChatGPT’s editing performance, my supervisor and I discussed possible strategies. One approach was to further modify the prompts by sharing our findings about the repeated vocabulary use with ChatGPT and asking it to replace some of these words. In addition, we considered specifying the prompts to better define ChatGPT’s role. For example, we could ask ChatGPT to edit the text as a violin pedagogue and scholar in the field of music education. By aligning ChatGPT’s role more closely with my own, it might select more relevant and appropriate vocabulary and expressions that better resemble my voice. Although my supervisor and I have not yet applied these strategies in practice, our discussion has directed us towards further exploring the ChatGPT-Human collaboration. This ongoing dialogue emphasizes the potential for refining AI prompts to enhance the alignment of ChatGPT’s output with the author’s intended voice and academic standards.

Supervisor: Reflections. As I read through the supervisee’s reflection about the beginning conversations that included utilizing ChatGPT, I was struck by the level of honesty and openness about her needs from a support perspective and an awareness of the time spent editing for those whose first language is not English. I continue to have great respect for those who write in a language that is not their first language. It also “spoke” to me in terms of how we as supervisors spend our time and whether it is the best use of time—not just for supervisors but for the supervisees. What might both of us be missing out on when our attention is focused on editing? The supervisee clearly identified this in her reflection. We began to discuss the differences, if any, between a professional editor and ChatGPT. And yes, while I was and continue to be suspicious and cautious, the supervisee thoughtfully convinced me to experiment within the parameters as outlined and articulated in her reflection. To offer support, I conversed with the Special IT advisor to the President on campus to walk through the processes as experienced and to seek insights from his expertise and the institutional perspective.

I was struck by the supervisee’s meta-analysis about her process when engaging with ChatGPT, but more importantly, how she clearly assumed a “teacher” role as she continually refined the prompts in an iterative manner. This required me to think about the value of utilizing ChatGPT.  The student’s clarity about the value of ChatGPT and how she utilized it contributed to my acceptance of its use. Finally, I learned about the importance of maintaining the author’s voice—in this case, the supervisee’s voice. We shared samples with the second reader, and the main point was, “Where was the student’s voice?”  This concern aligned with one of the ethical considerations we are addressing, that is, the supervisee’s voice.

This process was a learning curve for me. I had adopted the role of student as I listened to the supervisee as teacher while she explained to me how she was instructing ChatGPT. My role shifted back to supervisor when providing feedback to ensure that the supervisee’s voice remained present and intact, an ethical consideration, which is one focus of this paper. This shift in roles continued to be fluid as topics were introduced and discussions occurred. During the dialogical discourse, I experienced an increasing awareness of the supervisee’s voice as she continued in her writing process and as I continued to guide, respond to, and learn from her teaching, reflecting, and writing. I also noted the necessity to situate myself in a space of openness in which reflexivity required me to reflect on how and why the strengthening of I-Thou relationships is critical in educative engagements.

Question #2: In what Ways, if at all, do Roles of a Supervisee and Supervisor Shift during the Introduction of ChatGPT? How do the Shifts Impact, if at all, the Professional Relationships between Supervisee and Supervisor?

Supervisee: The role of a second language writer. In terms of how relationships are impacted as roles shift, and how my thinking grows through reflexivity and care, I plan to clarify my thoughts from two perspectives: the role of a second language writer and the role shifting between my supervisor and me.

As a non-native English speaker, my self-identity as a scholar in English-speaking academia has been significantly influenced by language barriers. In the field of music education, a high level of proficiency of English is essential. Differing from casual language use, academic writing demands a firm grasp of complex content and the ability to precisely articulate ideas, theories, data, and findings. In addition, formal tone and writing style are features of academic writing in which writers should avoid colloquial language, use precise vocabulary, and maintain a professional and an objective tone. Even native speakers struggle with academic writing, which highlights the additional challenges faced by second language scholars.

As aforementioned, working on academic writing can be time-consuming and frustrating. The process had undermined my confidence to be a scholar. As a novice researcher, my former supervisor spent considerable time editing and proofreading my text, focusing mainly on grammatical issues rather than my thoughts and research ideas. While I understood that addressing linguistic problems should be the first step before further discussions about content, I often felt frustrated and overwhelmed when receiving the edited file with over 200 comments. I could not think about where to start as it looked like a good-for-nothing patchwork. My intellectual contribution and distinctiveness were overshadowed by grammatical errors, leading me to interrogate myself: If I cannot conquer English academic writing, how can I be a qualified researcher? To be clear, this is not a reflection on the input of my former supervisor who provided many insights about my ideas but rather the reality of the supervisor’s role and their responsibility as an editor of students’ written works—for those whose first language and second language is English.

Supervisee: Role shifting between my supervisor and me. I agree that traditional academic writing training is essential for novice scholars, regardless of technological advancements. It is a challenging process for every researcher. However, my intention is not to judge the process as right or wrong. Instead, I want to honestly self-analyze how the traditional writing training has long suppressed my confidence. The appearance of ChatGPT, for researchers whose first language is not English, offers powerful academic writing assistance, and serves as a language learning tool.

My supervisor and I discussed the differences between a human editor and ChatGPT, assuming she had similar proofreading ability. So, why choose ChatGPT? Two reasons that emerge from the literature are the qualities of being objective and non-judgmental when generating AI responses. This could contribute to lessening overwhelming emotions in response to writing expectations, particularly for second language English writers.

In traditional supervisor-supervisee relationships, the supervisor is viewed as the expert in the academic field and in the dominant role. However, given that my supervisor and I decided to include ChatGPT to our collaborative work, I took the responsibility to introduce ChatGPT to her and kept her updated about my interaction with ChatGPT as she was not familiar with this technology in the beginning. Throughout the process of ChatGPT-Human collaboration, I took the main responsibility to prompt ChatGPT and adjust our interactions. This shifted the supervisor-supervisee relationship dynamic, making me the “expert” in our collaboration, as I transferred ChatGPT-related knowledge to my supervisor. Her role became that of a scrutinizer, examining my instructions to ChatGPT and evaluating its editing performance. Consequently, the traditional one-way knowledge transfer from teacher to student shifted into a mutual learning experience. Thus, our roles evolved from the conventional teacher-student dynamic to a more equal relationship, where both of us became learners and explorers in the ChatGPT-Human collaboration process.

Supervisor: Reflections. The supervisee’s reflections clearly focus on the roles and responsibilities of the supervisor and how our input could be more focused on ideas as the focus of the research project is investigated; supporting theoretical and methodological frameworks are devised; data are analyzed; and findings, discussions, and implications are presented. The question that consistently was entertained focused on how time is spent by the supervisor and supervisee—what the best use of time is in terms of moving the thinking forward.

I was struck by the supervisee’s reflections about her self-confidence as a thinker writing in a second language. A supervisor’s understanding and recognition of a supervisee’s ability to write in a second language is critical. What unchecked biases and assumptions are present as the supervisor begins the process of reading and evaluating the student’s work through the lens of knowing that supervisee identifies as a second language writer? And just as critical, how does that impact the student’s self-confidence, and thus one’s agency (one ethical consideration of this paper)? These unchecked biases and assumptions can impact pedagogical values of reflexivity and relational care as well as the place of dialogical discourse as each learn from the other.

It was also informative to read how she embraced the role of teacher as she taught and guided me through the capabilities of ChatGPT and identified my role as a “scrutinizer.” Notions of constructing what it means to shift in and through the roles of teacher, learner, and guide were reflected in our dialogical discourse. Interesting is the level of meta-cognitive reflection and reflexivity as the supervisee reflected on each of our roles, processes, and actions that occurred in response to her level of agency while utilizing ChatGPT. The supervisee clearly and succinctly identified how our roles shifted and the subsequent richness of those shifts from a pedagogical values perspective. I view this as an affirmation of relationship building, one built on trust, and our continuing discourse hopefully provided spaces in which my perception was affirmed.

Question #3: In what ways, if at all, does Utilizing ChatGPT as a Tool Impact the Development of a Supervisee’s Voice, Authorship, and Agency?

Supervisee: The authority and agency of Ph.D. learning and writing. The authority in Ph.D. learning and writing is embodied in the roles of supervisor and supervisee, implying dominant-subordinate relationship. In the traditional supervision process, the sense of dominance is difficult to be dissolved even if the supervisor is reflective and critical of their behaviors and expressions. The essence of authority is constructed by those in perceived positions of power, hence those who possess more knowledge and expertise are perceived to own stronger discourse authority (Foucault 1980).

This authority could be hidden and indiscernible. My supervisor and I had a conversation about how one might identify a graduate student’s supervisor based on the writing style. While one may believe that editing and proofreading behaviors are done in good faith to support students’ improvement, it cannot be denied that authority is subtly embedded in academic writing. When the third party, in this case ChatGPT, joins the traditional supervision process, the power dynamics starts to transform to a more balanced status. In other words, advanced generative AI possess information as human beings but provide responses and suggestions that are non-judgmental. To some degree, the involvement of ChatGPT diminishes traditional teacher’s authority as the main resource of knowledge. With the support from my supervisor, I was given more space to explore the effective way of learning and writing with ChatGPT, which in turn developed my ethos as a writer. Throughout the process, my sense of agency was amplified as I felt that I could contribute to the Ph.D. dissertation writing rather than merely waiting for instruction and receiving knowledge from my supervisor.

Supervisor: Reflections. The supervisee’s reflections required me to think through pedagogical values that are reflected in my practice. Questions that guided my thinking include what is my role? How do her reflections align with those whose writings have urged us to shift roles, focus on the critical and creative aspects of learning, engage with students’ agency so that voice and authorship are front and center because of an I-Thou relation? Her openness and understanding about that continual growth reflects a pedagogue who understands the key attribute of being an educator, that is, one who continues to learn and reflect, and learn upon reflection. It also affirmed my pedagogical value of working towards a relationship of care with the supervisee as her voice, authorship, and agency are not just recognized but independently nurtured. One new take away for me is how we as supervisors can work with students and self to identify one’s distinctive voice. This is critical as editors and readers of ChatGPT-generated editing responses. What is our voice and what is distinctive about the supervisee’s voice. If we are the editors as supervisors, how much is left as distinguished by the supervisee’s voice and that of the supervisor’s voice? Such reflective exercises can guide each of us as we examine ChatGPT’s editing comments and revisions. Only upon knowing the distinctiveness of one’s voice can we evaluate that which has been generated.

Question #4: In what ways, if at all, is one’s pedagogical values impacted as a result of discussions about and reflections on the above?

Supervisee: Lingering thoughts on pedagogical practice. The exploration on ChatGPT-Human collaboration with my supervisor has not only impacted my academic writing but has also impacted why one would utilize ChatGPT as a tool with continuing assessment through a critical lens. This experience prompted us to reconsider our pedagogies from a philosophical perspective. My supervisor’s open-mindedness to technologies with a relational care value has supported my exploration of personal learning and reflection on my role as a violin teacher. Questions I have been pondering include: In what ways can I create more space for younger generations to tailor their own playing? How can I make my support be most effective for different students’ needs? These questions will continue to guide my future research and guide my reflexive thinking as I reconstruct my pedagogical considerations. Genuine dialogues will always be crucial as I strive to understand students as independent human beings, and retrospectively scrutinize my own teaching practices.

Supervisor: Lingering thoughts on pedagogical practice. As I reflected on my pedagogical practice, I affirmed my values and necessary work of reflexivity and care, and of working with the student as she strengthens her voice, authorship, and agency while acknowledging the independent growth of each. What became apparent was the time spent editing students’ words without engaging her in the process, thus without practicing relational care. We discussed the criticality of me asking questions and engaging her in a dialogue about her writing process. We noted that the use of ChatGPT for very specific purposes in terms of editing could focus the interactions between supervisee and supervisor on dialoguing. During such discourse, we could question, reflect, and respond to further uncover what is meaningful to each other. This engagement could provide understanding about what is meaningful to the supervisee as she analyzes and interprets. This could deepen the supervisor’s and supervisee’s understanding about how voice, authorship, and agency are strengthened and as a value become part of one’s practice.

I continue to reflect on how shifting to a student role in which the supervisee guided me about how to teach ChatGPT was welcomed. It affirmed the notion that one of the greatest joys of being an educator is learning from those with whom you interact. It also reminded me of the criticality of growing while listening and responding with questions to further understanding.  As we, supervisee and supervisor, shift between roles of learning, facilitating, guiding, and teaching we can continue to examine our experiences through an educative lens (Dewey 1938) and reflect on the process of learnification (Guilherme 2019). This in turn can enrich I-Thou relation.

Conclusions

From the moment ChatGPT was involved in the doctoral dissertation editing process, the dialogue between us as supervisor and supervisee began to evolve. As the third party in the traditionally exclusive supervisor-supervisee relationship, ChatGPT redefined the supervisor’s authority as the sole expert transferring knowledge and editing the dissertation. Through collaboration with ChatGPT, the supervisee sometimes took the “teacher” role, guiding the AI, refining prompts, and sharing the experience of working with AI with the supervisor. In turn, the supervisor experienced a shift from the dominant teacher to a “scrutinizer,” carefully examining ChatGPT’s contributions and offering insights to improve the AI’s effectiveness. Both of us engaged in ongoing self-reflection and deep conversations, exploring the dynamics of our relationship. Framed through the lens of Buber’s theory of I-Thou, we came to see each other as subjects with independent and unique thoughts. Meaning emerged from the genuine dialogue between us. This fluid and dynamic role-shifting strengthened the supervisee’s agency. With ChatGPT as an assistant, the supervisee actively contributed to the editing process rather than passively waiting for instructions. The interaction enriched the supervisee’s role, transforming it into one of collaboration and co-creation rather than mere reception.

The dialogue between us continues as we write this paper together. Although it feels like a continuation of the collaborative process initiated during the doctoral dissertation editing, it is also something new, as different meanings emerge. Reflecting on what has driven the transformation in our supervisor-supervisee relationship, we realized that each of us brought unique perspectives to the dialogue and met each other where we were. For the supervisee, her voice and subjectivity were central. She sought to open an honest dialogue to express her struggles and concerns as a second-language writer in academia. Her voice was heard and accepted by the supervisor, who upheld the values of learning from others and offering care.

Much like puzzle pieces, it is only when the right pieces are placed in the right spots that they can fit together. Our perspectives, though distinct, are like two different puzzle pieces. It was our shared core approach of reflexivity that allowed these pieces to align and complete the picture. This alignment marked the transformation of our relationship into an I-Thou connection. No longer was I to you or I for you; instead, all that remained was I and you.

This paper, we hope, provides insights into the use of AI in music education and fosters dialogue for future research. We envision that such research may guide educators and learners as they continue to enter new realms of experiences in educational environments and consider roles and responsibilities; values of relational care, voice, authorship, and agency; and reflexivity as pedagogical values are examined.

About the Authors

Xiao Dong, Ph.D. in music education at Western University. Prior to her doctoral journey, Xiao received her master’s degree in violin performance at Soochow University, China. Xiao views her roles as a violinist, educator, and scholar as deeply intertwined. Her research interests have developed from her experience as a classically trained violinist. Xiao’s current research focuses on the development of metacognition and subjectification within music educational contexts. At the heart of her teaching philosophy lies the belief that true music education transcends the mere acquisition of knowledge and skills. Instead, it should foster a journey of self-discovery, inspiring transformative growth in learners.

Betty Anne Younker, (Adjunct Professor Emeritus), Dean of the Don Wright Faculty of Music at the University of Western Ontario from 2011-2021 and Professor of Music Education from 2011-2023. Previous appointments include The University of Michigan (2000-2011), University of Western Ontario (1997-2000) and University of Prince Edward Island (1992-1997). Her research has been published in national and international journals and as book chapters; and presented at national and international conferences. She has served in a variety of capacities including as President of the Michigan Music Educators Association, The College Music Society, and the London Arts Council; and as a member of several editorial boards and committees. Currently she serves on multiple CMS committees; Chairs the Board of Directors, Kiwanis Music Festival and Board of Directors, London Symphonia; is a member of the Publication Advisory Committee for CMEA; and is Past President for the University of Prince Edward Island Alumni Association.

References

Alvesson, Mats, and Sköldberg Kaj. 2017. Reflexive methodology: New vistas for qualitative research. Sage Publications.

Amirjalili, Forough, Masoud Neysani, and Ahmadreza Nikbakht. 2024. Exploring the boundaries of authorship: A comparative analysis of AI-generated text and human academic writing in English literature. Frontiers in Education 9 (March). https://doi.org/10.3389/feduc.2024.1347421

Baidoo-Anu, David, and Leticia Owusu Ansah. 2023. Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in prompting teaching and learning. Journal of AI 7 (1): 52–62. https://doi.org/10.61969/jai.1337500

Bakhtin, Mikhail. 1981. Discourse in the novel. In The Dialogic imagination, translated by Caryl Emerson and Michael Holquist, 259–422. University of Texas Press.

Benedict, Cathy. 2021. Music and social justice. Oxford University Press.

Bowman, Emma. 2022, December 19. A new AI chatbot might do your homework for you. But it’s still not an a+ student. NPR. https://www.npr.org/2022/12/19/1143912956/chatgpt-ai-chatbot-homework-academia

Brockmeier, Jens. 2009. Reaching for meaning: Human agency and the narrative imagination. Theory & Psychology 19 (2): 213–33. https://doi.org/10.1177/0959354309103540

Bruner, Jerome. 1986. Actual minds, possible worlds. Harvard University Press.

Bruner, Jerome. 1990. Acts of meaning. Harvard University Press.

Buber, Martin. 1958. I and thou. Translated by Ronald Gregor Smith. Charles Scribner’s Sons.

Camlin, David. 2015. “This is my truth, now tell me yours”: Emphasizing dialogue within participatory music. International Journal of Community Music 8 (3): 233–57. https://doi.org/10.1386/ijcm.8.3.233_1

Dergaa, Ismail, Karim Chamari, Piotr Zmijewski, and Helmi Ben Saad. 2023. From human writing to artificial intelligence generated text: Examining the prospects and potential threats of ChatGPT in academic writing. Biology of Sport 40 (2): 615–22. https://doi.org/10.5114/biolsport.2023.125623

Dewey, John. 1938. Experience and education. The Kappa Delta Pi Lecture Series. Collier Books, Macmillan Publishing Company.

Fahimirad, Mehrnaz, and Sedigheh Shakib Kotamjani. 2018. A review on application of artificial intelligence in teaching and learning in educational contexts. International Journal of Learning and Development 8 (4): 106–18. https://doi.org/10.5296/ijld.v8i4.14057

Foucault, Michel. 1980. Power/knowledge: Selected interviews and other writings 1972–77. Edited and translated by Colin Gordon. Pantheon Books. http://ci.nii.ac.jp/ncid/BA05110939.

Freire, Paulo. 1970. Pedagogy of the oppressed. 2nd ed. Penguin.

Grassini, Simone. 2023. Shaping the future of education: Exploring the potential and consequences of AI and ChatGPT in educational settings. Education Sciences 13 (7): 692–705. https://doi.org/10.3390/educsci13070692

Guilherme, Alex. 2019. AI and education: The importance of teacher and student relations. AI & Society 34 (1): 47–54. https://doi.org/10.1007/s00146-017-0693-8.

Helms-Park, Rena, and Paul Stapleton. 2003. Questioning the importance of individualized voice in undergraduate L2 argumentative writing: an empirical study with pedagogical implications. Journal of Second Language Writing 12: 245–65. https://doi.org/10.1016/j.jslw.2003.08.001

Hess, Juliet. 2021. “Putting a face on it”: The trouble with storytelling for social justice in music education. Philosophy of Music Education Review 19 (1): 67–89. https://doi.org/10.2979/philmusieducrevi.29.1.05

Hibbard, Shannan. 2017. Music teacher presence: Toward a relational understanding. PhD diss., University of Michigan.

Hibbard, Shannan. 2021. Disrupting “what we know too well”: A relational frame for considering trauma in music education. In Trauma and resilience in music education, edited by Deborah Bradley and Juliet Hess, 35–48. Routledge.

Holzkamp, Klaus. 1983. Grundlegung der psychologie (Foundations of psychology). Campus Verlag.

Maphoto, Kgabo Bridget, Kershnee Sevnarayan, Ntshimane Elphas Mohale, Zuleika Suliman, Tumelo Jacquiline Ntsopi, and Douglas Mokoena. 2024. Advancing students’ academic excellence in distance education: Exploring the potential of generative AI integration to improve academic writing skills. Open Praxis 16 (2): 142–59. https://doi.org/10.55982/openpraxis.16.2.649.

Mitchell, Elizabeth, and Cathy Benedict. 2020. Lives in dialogue: Shared musical-relational engagements in music therapy and music education. European Journal of Philosophy in Arts Education 5 (1): 33–67. https://doi.org/10.5281/zenodo.3935805

Mondal, Himel, and Shaikat Mondal. 2023. ChatGPT in academic writing: Maximizing its benefits and minimizing the risks. Indian Journal of Ophthalmology/Indian Journal of Ophthalmology 71 (12): 3600–06. https://doi.org/10.4103/ijo.ijo_718_23

Nguyen, Andy, Yvonne Hong, Belle Dang, and Xiaoshan Huang. 2024. Human-AI collaboration patterns in AI-assisted academic writing. Studies in Higher Education (February): 1–18. https://doi.org/10.1080/03075079.2024.2323593

Noddings, Nel. 2005. Caring in education, The Encyclopedia of pedagogy and informal education. https://infed.org/mobi/caring-in-education/

Semrl, Neli, Feigl S, Taumberger Nadja, Bracic T, Fluhr H, Blockeel C, and Kollmann M. 2023. AI language models in human reproduction research: Exploring ChatGPT’s potential to assist academic writing. Human Reproduction 38 (12): 2281–88. https://doi.org/10.1093/humrep/dead207

Smith, Mark. 2012. What is pedagogy? A definition and discussion. The Encyclopedia of Pedagogy and Informal Education. https://infed.org/mobi/what-is-pedagogy/

UNESCO. 2021. AI and education: Guidance for policy-makers. UNESCO. https://doi.org/10.54675/pcsp7350

Wise, Beck, Lisa Emerson, Ariella Van Luyn, Bronwen Dyson, Collin Bjork, and Susan E. Thomas. 2024. A scholarly dialogue: Writing scholarship, authorship, academic integrity and the challenges of AI. Higher Education Research & Development/Higher Education Research and Development 43 (3): 578–90. https://doi.org/10.1080/07294360.2023.2280195