Volume 8 | Nomor 4 | Tahun 2025 | Halaman 1093Ai1104
E-ISSN 2615-8655 | P-ISSN 2615-725X
http://diglosiaunmul.
com/index.
php/diglosia/article/view/1460 Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students Negosiasi identitas, keaslian, dan kemandirian mahasiswa Pendidikan Bahasa Inggris dalam penulisan proposal disertasi berbantuan kecerdasan buatan Widhiyanto1,*.
Rahayu Puji Haryanti2.
Ira Mutiaraningrum3.
Candradewi Wahyu Anggraeni4.
Lilis Suryani5, & Mohammad Ali Yafi6 Universitas Negeri Semarang Sekaran.
Gunungpati.
Semarang.
Indonesia Politeknik Negeri Sambas Kawasan Pendidikan Tinggi.
Jl.
Sejangkung.
Sambas.
Indonesia Universitas Tidar Jl.
Kapten Suparman.
Magelang.
Indonesia IKIP Siliwangi Jl.
Terusan Jenderal Sudirman.
Cimahi.
Indonesia Universitas Tunas Pembangunan Jl.
Balekambang Lor.
Surakarta.
Indonesia *Email: pakwidhi_english@mail.
Orcid: https://orcid.
org/0000-0001-7351-8976 Email: rahayu_ph@mail.
Orcid: https://orcid.
org/0009-0005-6393-6011 Email: ipoltesa@gmail.
Orcid: https://orcid.
org/0000-0003-2273-5203 Email: candradewi@untidar.
Orcid: https://orcid.
org/0000-0003-1610-1124 Email: lis-suryani@ikipsiliwangi.
Orcid: https://orcid.
org/0009-0004-6367-486X Email: mohammadaliyafi@lecture.
Orcid: https://orcid.
org/0000-0002-6942-9791 Article History Received 23 August 2025 Revised 3 October 2025 Accepted 16 October 2025 Published 26 December 2025 Keywords dissertation proposal Kata Kunci kecerdasan buatan.
penulisan proposal Read online Scan this QR code with your smart phone or mobile device to read Abstract The use of AI as a writing assistant among doctoral students is an emerging practice.
There has not been much research exploring this topic.
Previous studies on the use of AI for final projects have primarily focused on language learning, despite this genre having unique academic requirements, such as academic identity, scientific integrity, and agency.
This study examined the diverse approaches doctoral students take to developing an academic voice when utilizing AI.
It examines the manifestation of academic integrity and agency, as well as the interplay among these constructs, through a case study.
Thematic analysis, as a data analysis tool, from the responses of six doctoral students enrolled in the Proposal Dissertation Course, sheds light on the critical acceptance of AI.
The results of the focus-group discussion and studentsAo reflective journals reveal that AI is recognized as applicable but does not replace human critical thinking in the manual investigation process.
Research shows that students build academic identity through strategies such as making AI a discussion partner to maintain the authorAos voice.
They also maintain academic integrity by verifying AI sources, avoiding unreal citations, and manually paraphrasing the text.
Students leverage AI for technical tasks and time management yet remain non-reliant and edit AI output to ensure it remains original work.
Abstrak Penggunaan kecerdasan buatan (AI) sebagai asisten penulisan di kalangan mahasiswa program doktoral merupakan praktik yang sedang berkembang luas di seluruh dunia.
Namun, belum banyak penelitian yang mengeksplorasi fenomena ini.
Penelitian ini mengeksplorasi pendekatan mahasiswa program doktoral dalam membangun suara akademik saat menggunakan AI dalam menulis tugas akhir.
Penelitian ini menganalisis manifestasi integritas, identitas, dan otonomi akademik, serta interaksi di antara ketiga konstruk tersebut dalam proses penulisan proposal disertasi mahasiswa Pendidikan Bahasa Inggris.
Dengan melibatkan enam mahasiswa doktoral yang mengikuti Mata Kuliah Proposal Disertasi, penelitian ini menyoroti penerimaan kritis terhadap AI.
Penelitian ini menggunakan desain studi kasus.
Data yang terkumpul lewat wawancara dianalisis dengan analisis tema.
Hasil diskusi kelompok terfokus dan jurnal reflektif mahasiswa menunjukkan bahwa kecerdasan buatan (AI) diakui sebagai alat yang berguna, namun tidak menggantikan pemikiran kritis manusia melalui proses penyelidikan manual.
Penelitian menunjukkan bahwa mahasiswa membangun identitas akademik melalui strategi seperti menjadikan AI sebagai mitra diskusi untuk mempertahankan suara penulis.
Mereka juga menjaga integritas akademik dengan memverifikasi sumber AI, menghindari kutipan yang tidak valid, dan mengutip teks secara manual.
Siswa memanfaatkan AI untuk tugas teknis dan manajemen waktu, namun tetap tidak bergantung padanya dan mengedit jawaban AI untuk memastikan keaslian karya mereka.
A 2025 The Author.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya by Universitas Mulawarman How to cite this article with APA style 7th ed.
Widhiyanto.
Haryanti.
Mutiaraningrum.
Anggraeni.
Suryani.
, & Yafi.
Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya, 8.
, 1093Ae1108.
https://doi.
org/10.
30872/diglosia.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya is an open access article under the terms of the Creative Commons Attribution-Share Alike 0 International License (CC BY-SA 4.
Widhiyanto.
Rahayu Puji Haryanti.
Ira Mutiaraningrum.
Candradewi Wahyu Anggraeni.
Lilis Suryani, & Mohammad Ali Yafi Introduction Over the last decade, artificial intelligence (AI, hereafte.
has revolutionized numerous aspects of human life, including how people think and work in academic fields.
Platforms like ChatGPT.
Grammarly, and Claude are now utilized in brainstorming, drafting, and final editing.
These platforms have been reported to shift the traditional role of writing as a critical thinking process to a machine-facilitated technical activity (Gabriel, 2024.
Song, 2.
Several studies have shown that technology has created new efficiencies in academic text production, but it simultaneously raises concerns about the loss of authenticity and reflection in writing (Werdiningsih et al.
, 2024.
Wiredu et al.
, 2.
When students can produce scientific text in a much shorter time, old values such as depth of analysis and epistemic understanding begin to be marginalized.
This condition creates new pressure for educational institutions to redefine the goals of academic literacy in the era of advanced technology.
Several universities have even developed official policies to guide students in AIAos ethical and reflective use (Marcel & Kang, 2.
Furthermore, the use of AI in academic writing has disrupted the traditional thinking process and writing structure.
Generative AI is capable of constructing logical arguments, summarizing literature, and even paraphrasing in an academic style, so that students do not need to undergo the deep thinking process that usually accompanies writing a dissertation proposal (Taiye et al.
, 2.
Cognitive processes previously inherent in writing activities are now replaced by automated capabilities of large language models (LLM.
, which indirectly reduce studentsAo mental engagement with the content they write (Jelson et al.
, 2025.
Jungherr, 2.
Study shows that many students prefer to use AI to avoid confusion when starting to write, but this reduces their potential to learn from the process of making mistakes and revising (Wang et al.
, 2.
Additionally.
AI tends to produce uniform text structures that lack personal logical development, thereby reducing stylistic and structural differentiation among students.
In this context, disruption is not merely about speed but touches on the fundamental aspects of academic epistemology.
More concerning, the existence of AI also challenges the traditional values that have long been the foundation of academic or scientific writing.
Once defined as the result of individual contemplation and exploration, originality is now reduced to the ability to direct the right prompt (Adauto Medina et al.
, 2024.
Gonsalves, 2.
Similarly, critical reflection, which previously marked academic maturity, can be substituted by textual output that appears logical but lacks depth of interpretation (Timmis et al.
, 2.
Furthermore, mastery of academic discourse, which is the result of prolonged interaction between writers and specific intellectual traditions, is also threatened by the ease of technology (Sarkar, 2.
In this situation, the question arises: Are students still forming their academic voices, or are they merely becoming technology users who produce content with little understanding of its content and context? Therefore, educational institutions need to introduce technology and instill critical and ethical principles in its use (Bozkurt, 2.
The following sections delve into studentsAo academic identity, integrity, and agency during AI use for academic purposes.
StudentsAo academic identity is a social and cognitive construct formed through the interactional process, reflection, and self-representation in the academic community.
This identity is not a fixed trait, but rather one that is formed through discourse and academic norms developed over time within an institutional environment (Lea & Street, 1.
Students construct this identity through the mastery of academic genres, use of scientific language, and affiliation with specific scientific values, which indicate their position in the epistemic community (Brown, 2.
In this context, writing activities, especially dissertation proposals, become crucial opportunities to show how students position themselves as researchers and contributors in their respective fields.
This process demands technical competence in writing, self-reflection, argumentation formation, and understanding of the academic audience (Grigorenko, 2.
This academic identity develops through the active involvement of students in selecting literature, formulating research questions, and developing theoretical frameworks rooted in their understanding of their discipline (Butarbutar & Gonzylez Vallejo, 2.
The presence of AI has altered the dynamics of this academic identity.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students as some cognitive processes have been mediated or even replaced by algorithms.
Students who use AI intensively may lose the opportunity to experience the crucial stages of deep learning that shape their academic skills (Parker et al.
, 2.
Dependence on AI will also decrease studentsAo intellectual independence in sorting literature, arranging arguments, and addressing the complexity of research methodology (Luo et al.
, 2.
As a result, studentsAo academic voices have become weak because most emerging narratives result from algorithmic calculations rather than original personal reflections (Pasenta & Chakim, 2.
Next, academic integrity serves as the primary foundation for addressing the ethical challenges of AI in dissertation proposal writing.
Integrity is reflected in studentsAo efforts to understand, process, and convey ideas in depicting cognitive work and ethics.
When AI is involved in writing, the border between technical assistance and intellectual abdication becomes blurred.
Without this strong ethical understanding, students risk violating the principle of originality by adopting AI outputs raw, without undergoing a critical evaluation process (Carobene et al.
, 2.
The temptation of efficiency offered by AI often encourages students to skip substantial analysis stages that should strengthen their scientific capacity.
Some students use AI to compile literature reviews, formulate conceptual frameworks, and even synthesize findings automatically.
Therefore, many institutions have begun to formulate policies that emphasize the importance of disclosing AI use, limitations on its intervention, and early training in AI ethics literacy for students (Wiredu et al.
, 2.
Integrating AI principles into academic writing curricula is expected to strike a balance between the use of technology and strengthen studentsAo independent scientific capabilities (Lund et al.
, 2.
Lastly, student agency is crucial in AI-assisted writing.
Student agency is characterized as the ability of students to actively shape their learning experiences, make informed decisions, and manage their own learning.
The concept of agency correlates with behavioral and motivational aspects, emphasizing studentsAo ability to govern their actions, make informed decisions, and navigate complex social environments (Adie et al.
, 2018.
Darvishi et.
al Students play a crucial role in utilizing AI wisely and effectively.
As we know.
AI has a wide range of applications, including in the teaching and learning process.
Students are responsible for the writing process, assisted by AI.
They not only accept all the content from AI but also criticize it.
Thus, agency refers to studentsAo capacity to make reflective decisions and assume responsibility.
Agency connects identity and integrity: a stronger agency correlates with a more authentic academic identity and enhanced The studentAos agency emphasized the responsibility of the educational role.
Here are some experts who integrate studentsAo agency into AI.
Black and Tomlinson .
explain that AI is vital for improving their work in writing, but many students expressed skepticism about AI-generated They also distinguished between AI-assisted editing and their original thinking.
The students realized that AI is beneficial, but they do not fully believe in all its content.
So, they criticise and rethink the ideas that AI provides.
This statement shows studentsAo responsibility for using AI.
Another study shows that while AI tools make things faster and more accurate, overusing them may make it harder to learn to think critically, be creative, and write independently.
Besides, excessive use decreased cognitive effort and creativity, made personal writing styles less critical, and reduced confidence and skill retention (Budiyono et.
al, 2.
Previous studies have often discussed academic integrity .
, plagiarism and AI detectio.
or separately explored student agency and academic identity.
Almost none have combined these three dimensions to examine how they interact with each other when students use AI.
The experiences of students in this study are influenced by different academic settings, rules, and cultures, such as norms regarding the use of AI, relationships with lecturers, and perceptions of originality, which reinforces the uniqueness of this study.
This provides important contextual data that may not have been widely reported.
In addition, previous studies have not focused on the dissertation proposal, where students develop their research framework and vision, so this study provides insights into this important early stage of academia.
AI creates a new thing: students must determine the extent to which they represent their identities, maintain the authenticity of their ideas, and manage intelligent tools ethically and strategically.
This research aims to explain how agencies assess the Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Widhiyanto.
Rahayu Puji Haryanti.
Ira Mutiaraningrum.
Candradewi Wahyu Anggraeni.
Lilis Suryani, & Mohammad Ali Yafi quality and authenticity of studentsAo academic contributions in the era of AI.
As a result, the research questions of the study are: .
how do students demonstrate diverse approaches to constructing an academic voice when using artificial intelligence? .
what are the manifestations of academic integrity in studentsAo use of artificial intelligence? .
what role does student agency play in the use of artificial intelligence for academic purposes? .
how do identity, integrity, and agency dynamically interact in studentsAo engagement with artificial intelligence? Method This study uses a case study approach.
A case study is appropriate for this study because it examines a specific context: ELT students writing dissertation proposals with the help of AI.
Findings related to studentsAo personal experiences, perceptions, and strategies cannot be separated from their educational setting.
The data used in this study were reflective journals and interviews.
Reflective journals were collected via Google Forms, while interviews were conducted through focus group discussions via Zoom.
The interview lasted 1 hour and 50 minutes.
ParticipantsAo original quotes .
oded Respondent 1-.
were the primary evidence supporting the interpretation.
We maintained the authentic voice of the students without fabricating data.
This study included six respondents in their second semester at one of IndonesiaAos public They were enrolled in the Dissertation Proposal Writing Course.
The six respondents had varying degrees of teaching experience and AI use.
Some had been teaching for between 5 and 16 years, with AI usage ranging from 1 to 2 years.
The AI tools used included ChatGPT.
Grammarly.
Quillbot.
Paperpal.
Jenni, and Gemini, while one respondent had never used AI.
Before the study, the students were given a letter of consent to become research respondents.
The faculty has approved this letter as it is part of the universityAos research program.
The letter also explained that their data would be kept confidential and used only for research purposes.
The study also used codes Respondent 1-Respondent 6 to maintain anonymity.
The data analysis in this study uses thematic analysis.
The research results are presented in a table containing the themes and subthemes of the research findings.
The thematic analysis in this study includes data familiarization, code generation, theme construction, theme definition, and the final report (Braun & Clark, 2.
We ensure the validity of research data by utilizing various data sources, including interviews and studentsAo reflective journals.
This helps ensure that the results are not biased toward one type of data.
We also conduct peer debriefings to ensure that the analysis results are not solely based on a single interpretation.
We conducted an iterative analysis by comparing new codes and themes with existing ones to identify any new codes or patterns that We considered data saturation when additional data no longer produced new codes or In qualitative analysis, particularly thematic analysis, presenting results in the form of a table of themes and subthemes is used to summarize findings in a structured and easily readable manner.
Tables enable researchers to display hierarchical relationships between main themes and subthemes, while also presenting examples of raw data, such as respondent quotes, thereby making the analysis process transparent and traceable.
We present the original interview quotes within the body to provide richer context and a more flowing narrative.
Findings and Discussion Academic Identity in the Age of AI This present study shows that students demonstrate various ways in shaping academic voice when they utilize AI technology in writing dissertation proposals.
Some students use AI wisely as a discussion partner to help them formulate initial ideas, while still maintaining their own Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students argumentative structure and writing style.
They demonstrate an awareness not to surrender complete control of writing to technology, but to make AI a collaborative partner.
AuChatGPT is quite helpful for discussing strange academic things.
Google is sometimes less talkative, less Ay (Respondent 4, 51:10Ae53:.
AuWhen GPT provides answers.
I didnAot use it entirelyA I considered which ones to use, matching them with what I want.
Ay (Respondent 5, 48:07Ae48:.
Table 1.
The Way Students Construct Academic Identity Through the Use of AI during Dissertation Proposal Writing Theme Strategies in Constructing Academic Voice as Identity with Artificial Intelligence (AI) Sub-theme AI as a Human-Like Discussion Partner.
Balancing AI Output with Researcher Intentions.
AuthorAos Voice as the Core of Originality.
Students build their academic identity using AI in writing dissertation proposals.
Students reveal that AI functions as a discussion partner with whom they can engage in dialogue.
They feel that AI helps them clarify their ideas and arguments.
Students in this study balance the output of AI with their own intentions so that the results are not merely technical or generic.
They maintain the AoauthorAos voiceAo to preserve originality.
In this study, students continue to make decisions, choose language styles, and construct arguments according to their personal identities.
This ensures that their writing continues to reflect their voice and identity.
This approach shows their effort in balancing their AI outputs.
It depicts active negotiation between the efficiency of technology and the authentic expression of academic identity.
These findings are consistent with Grigorenko .
and Butarbutar and Gonzylez Vallejo .
, who state that academic identity is formed through active engagement in the process of thinking and Conversely, students who rely too much on AI tend to miss out on the opportunity to experience the deep cognitive processes that are essential for building their academic character (Luo et al.
, 2024.
Parker et al.
, 2.
Thus, academic identity becomes an arena of negotiation between self-authenticity as an academic and the efficiency offered by generative technology (Taiye et al.
, 2.
Manifestation of Academic Integrity in Using AI for Dissertation Proposal Writing Academic integrity in the context of AI is understood as a construct that depends on studentsAo understanding, values, and experiences.
Students with high ethical awareness tend to openly declare their use of AI in their writings and critically review the generative results before using them as a final draft.
This practice embodies a form of intellectual responsibility and commitment to originality, a key principle in academia (Haber et al.
, 2.
Therefore, academic integrity is not merely compliance with institutional regulations, but a personal commitment to scientific honesty and reflective thinking processes.
When academic integrity is tested by AI efficiency, studentsAo decisions to use or edit AI results become a key indicator of their ethical stance.
Students with clear values are less likely to succumb to the temptation of shortcuts but instead use AI to strengthen arguments they have designed They engage deeply, thoughtfully, critically, and independently with complex texts (Pujiastuti et al.
, 2.
The ability to select information to be included in a written narrative demonstrates studentsAo cognitive abilities (Andriani et al.
, 2.
The ability to construct arguments critically is a 21st-century skill that students need (Boeriswati et al.
, 2.
Students in this study were aware that some references in AI might be nonexistent.
They also checked their writing using Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Widhiyanto.
Rahayu Puji Haryanti.
Ira Mutiaraningrum.
Candradewi Wahyu Anggraeni.
Lilis Suryani, & Mohammad Ali Yafi Turnitin to detect the AI percentage and performed manual paraphrasing to reduce it.
Here are some passages from studentsAo interviews:
AuManual paraphrasingA I read it again, retyped it, and revised it so it wouldnAot be detected as AIgeneratedA I was worried it might even increase if I used AI for paraphrasing.
Ay (Respondent 6, 54:55Ae 55:.
AuI didnAot sleep for two days trying to lower the AI percentageA I used a thesaurus to find synonyms and rewrote it again.
Ay (Respondent 2, 56:35Ae58:.
Table 2.
StudentsAo Academic Integrity during Dissertation Proposal Writing Theme Academic Integrity in the Use of Artificial Intelligence (AI) Sub-theme Cross-check and validate AI-generated sources Avoid citing non-existent sources and ensure that all references are drawn from verifiable, published works Intensive Effort to Lower AI Percentage.
Manual Paraphrasing to Reduce AI Detection Students maintain academic integrity in writing their dissertation proposals by cross-checking and validating sources generated by AI.
They ensure that the sources provided by AI exist and download them.
They also ensure that the percentage of AI usage is reduced through manual paraphrasing, making their writing more original.
The findings align with Lund et al.
, which highlight the importance of integrating AI ethical principles into academic writing curricula to ensure students are not only technically skilled but also morally resilient.
Academic integrity extends beyond just an administrative assessment.
is a manifestation of studentsAo critical awareness in interacting with technology in an ethically responsible manner.
These findings reinforce the views of Bozkurt .
and Gabriel .
that higher education should prioritize deep technological ethics, not just superficial regulations.
The Role of Student Agency in AI Use Based on the findings, studentsAo role as a student agency is crucial in using AI, especially in Agency is a key determinant of how students use AI reflectively or passively.
Students in this study demonstrated their agency by planning their use of AI before writing.
They utilize AI for technical tasks such as formatting .
dhering to the correct style guid.
, data presentation .
reating and labeling tables, figures, and chart.
, consistency checks .
nsuring consistent terminology, units, and abbreviation.
, structural organization .
rranging content logicall.
, and technical proofreading .
hecking grammar and punctuatio.
Students can sort AI functions according to needs, adapt AI suggestions contextually, and redevelop ideas with critical thinking.
The students have a role in editing, criticizing, and developing their ideas.
Students with high agency can utilize AI prompts as guidance, developing self-regulation strategies, and maintaining their voice and authenticity (Darvishi et al.
, 2024.
Wang.
The students also utilize AI to enhance the effectiveness and efficiency of their writing without compromising their academic tone (Ginting et al.
, 2.
that hinders them from being overly reliant on AI.
What follows are excerpts drawn from interviews with students AuI used AI not because I was lazy.
I tried to finish quickly and made AI generate ideas for me.
Ay (Respondent 1, 01:08:18Ae01:09:.
AuIf itAos purely 100% by AI.
I think thereAosA we need to know the limitsA when we have to slow down and when we can go ahead.
Ay (Respondent 3, 01:13:02Ae01:14:.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students AuI ended up writing it myselfA the theory from my own ideas, from articles I readA the problem is, my own writing was detected as AI-generated.
Ay (Responden 4, 01:08:18Ae01:09:.
Table 3.
StudentsAo Agency in Writing Theme Learning agency in academic writing supported by AI Sub-theme Organizing writing plan before using AI Use AI for technical writing tasks AI for Time Management Avoiding Over-Reliance on AI Editing AI Output to Retain Ownership By using AI in writing dissertation proposals, students gain learning agency in academic writing, as they develop a writing plan before utilizing AI.
AI is used for technical tasks such as checking grammar and format, as well as helping with time management to make the writing process more efficient.
On the other hand, the respondents in this study are aware that they should not rely on AI.
This is evident from the interview results, which show that they still edit the answers provided by AI.
Students discovered that their original writing was flagged as AI-generated.
They paid attention to the AI checker and Turnitin percentage during the proposal writing process.
This indicates that they have a vital role in the context of their writing.
Integrity is essential for doctoral students since AI may hinder student initiative and creative risk-taking (Li & Wilson, 2025.
Black & Tomlinson, 2.
Dynamic relationship: Identity-Integrity-Agency These three concepts interact with and form a technology-based academic learning ecosystem.
Students must realize their identity and sharpen their integrity to have high agency toward the responsibility of using AI in a writing context.
A strong academic identity develops in students with high agency and a commitment to integrity.
The students who know their identity as academic persons realize they must be responsible for who they are.
They also have complete responsibility for using AI in writing contexts.
The research suggests that studentsAo academic identity influences their self-regulation and commitment, leading to improved educational outcomes and personal growth (Human-Vogel.
The students also commit to ethical academic behavior and contribute to a positive learning environment (Brickhill et.
, 2.
Integrating AI as a support, not a replacement, ensures that students remain in control of the academic process.
The authenticity of research is maintained when AI assistance is balanced with the active involvement of researchers in developing ideas and verifying information (Chen et al.
, 2.
, thereby strengthening, rather than replacing, the role of Below are selected excerpts from student interviews:
As educators .
we must evaluate students so that they do not rely entirely on AI.
giving verbal tests, allowing the originality of their ideas to become visible.
Ay (Respondent 4, 01:15:39Ae01:16:.
AuAI is the new normalA not a disgraceA the important thing is how we use it.
Ay (Speaker 0, 01:13:02Ae 01:14:.
Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Widhiyanto.
Rahayu Puji Haryanti.
Ira Mutiaraningrum.
Candradewi Wahyu Anggraeni.
Lilis Suryani, & Mohammad Ali Yafi Table 4.
Interplay Among Identity.
Integrity, and Agency during AI Use Theme Interplay Among Identity.
Integrity, and Agency in AI Use Emotional and Perceptual Responses to AI in Academic Work Normalization and Ethical Responsibility in AI Integration Sub-theme Integrating AI as a support, not a replacement Balance between AI Assistance and Research Authenticity Mixed Feelings: Confidence vs.
Anxiety AI Detection as a Source of Skepticism AI as a Normalized Academic Tool Responsibility of Educators to Address AI Use Students express mixed feelings of confidence and anxiety when using AI in an academic One source of anxiety is the existence of AI detection systems that cast doubt on the originality of their work.
Their reputations could be damaged if AI detection tools accuse them of being dishonest or fraudulent due to inaccurate AI detection results (Giray et al.
, 2.
Nevertheless.
AI is beginning to be viewed as a normalized academic tool and part of the daily learning process.
In this situation, educators hold an important responsibility to guide the ethical and productive use of AI.
AI can be a tool to strengthen oneAos identity when used consciously .
hrough agenc.
and ethically .
ith integrit.
AI can help students organize their ideas, provide feedback, and enable them to explore the writing context.
How students are responsible for using AI can shape their AI does not replace the role of students, but it helps them write effectively without compromising their authenticity in the context of writing (Nugroho et al.
, 2024.
Although AI makes education more accessible and practical, the intentional misuse of AI systems in promoting academic cheating has become a significant concern (Agha, 2025, as cited in Oyedokun.
StudentsAo responsibility must be emphasized to ensure the authenticity of studentsAo writing.
Table 4 presents the dynamic relations among identity, integrity, and agency.
When agency and integrity are compromised.
AI can undermine the thinking process and hinder authentic academic expression.
Students are passive when they use AI for academic writing, without criticizing and developing their ideas.
They lack practical critical thinking and rely solely on AI.
Dependency on AI can reduce studentsAo critical thinking and threaten the authenticity of their writing (Octaberlina et al.
, 2.
Therefore, strengthening studentsAo commitment to academic integrity in the AI era is necessary to address the challenges of using AI (Benke & Szcke, 2.
should be used as a partner, not as a replacement for studentsAo critical thinking and authentic voices.
Conclusion This study aims to explore the relationship between academic identity, academic agency, and academic integrity in doctoral studentsAo dissertation writing.
The findings highlight that students demonstrate diverse ways of constructing an academic voice when using artificial intelligence, manifestations of academic integrity, the role of student agency in artificial intelligence use, and dynamic relationships among identity, integrity, and agency.
AI can be a tool to strengthen oneAos identity when used consciously .
hrough agenc.
and ethically .
ith integrit.
The studyAos findings pave the way for designing an academic writing curriculum that teaches not only writing conventions but also awareness of oneAos identity as an academic.
Artificial Intelligence literacy must encompass both technical and ethical dimensions, instilling values of reflection, responsibility, and intellectual autonomy.
It is recommended that higher education institutions should foster spaces for critical discourse on technology ethics, rather than merely imposing This study has limitations, including a small sample size and being conducted in a single class within a single course, where the influence of the institution, class culture, or lecturer can be particularly pronounced and therefore not representative of a broader population.
Self-reports also have the potential for bias due to the respondentsAo memories and the conditions at the time of the interview, as well as the researchersAo interpretation of the data analysis.
Furthermore, this study Diglosia: Jurnal Kajian Bahasa.
Sastra, dan Pengajarannya Vol.
8 No.
1093Ai1104 Negotiating identity, originality, and agency in AI-assisted dissertation proposal writing among English education students only evaluates the proposal writing stage, rather than tracking how studentsAo identity, originality, and agency change from the beginning to the end of their dissertations.
Future research could involve direct observation, document analysis .
uch as dissertation proposal draft.
, or examining changes in studentsAo writing before and after using AI.
Acknowledgements The authors would like to extend their Gratitude to Universitas Negeri Semarang for sponsoring this research through Penelitian Dasar DPA LPPM UNNES 2025 Number 3/UN37/PPK.
11/2025.
References