From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students

INTERNATIONAL JOURNAL OF MULTIDISCIPLINARY RESEARCH AND ANALYSIS ISSN(print): 2643-9840, ISSN(online): 2643-9875 Volume 08 Issue 09 September 2025 DOI: 10.47191/ijmra/v8-i09-52, Impact Factor: 8.266 Page No. 5390-5398 IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5390 From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students Ceasar Ian P. Sabillo, PhD1, Priscilla A. Marayag, PhD2 1Glan Institute of Technology Glan, Sarangani Province, Region XII, Philippines, 9517 2Dean, College of Teacher Education GenSantos Foundation College, Inc.General Santos City, Philippines, 9500 ABSTRACT: This qualitative phenomenological study examined the experiences of 3rd-year BSTM students at Glan Institute of Technology, Philippines, concerning AI integration in thesis development. The findings indicate that AI enhances efficiency in literature reviews, writing refinement, and data analysis; however, it encounters significant barriers, including unreliable infrastructure, financial constraints, and output limitations. Ethical concerns include diminished originality, superficial engagement and algorithmic bias. The study concludes that while AI provides substantial support, its responsible adoption necessitates institutional interventions to ensure equitable access, incorporate critical AI literacy training, and redesign assessments to prioritize process transparency over output expediency, thereby upholding academic integrity and authentic scholarship. KEYWORDS: Artificial Intelligence (AI), Thesis Development, Integrity, Critical Literacy INTRODUCTION In the contemporary era of technology, the incorporation of artificial intelligence (AI) has considerably transformed multiple sectors, including research. One noteworthy domain in which AI has made a substantial impression is scholarly paper production. As a result of progress in AI innovation, researchers can now utilize potent resources that aid them at various points in the research process. Advancements in artificial intelligence have led to the development of innovative electronic writing tools, several of which are novel Alharbi (2023). These writing support systems are useful for writers, especially students, because they provide suggestions for sentence completion and text generation assistance during and after the writing process. While the adoption of these tools by students has resulted in an increase in scientific publications on the subject, the findings of these studies are often conflicting, and their credibility may be questioned. Artificial intelligence has transformed the thesis-writing process by providing students with efficient and effective tools for research and writing. By leveraging AI capabilities, students can easily streamline data collection, analysis, and synthesis, thereby making the entire process less challenging. The integration of AI into academic writing has significantly improved the quality of research and writing, allowing students to produce exceptional work with minimal effort and time. Nazari et al. (2021) identified students’ affective characteristics as a significant challenge in educational technology integration. Additionally, research on the impact of technology on attitudes and learning behavior is limited. AI-powered writing tools can enhance learning behavior and attitudinal technology acceptance by providing formative feedback and assessment to non-native postgraduate students in English academic writing. Using artificial intelligence can greatly improve the quality of undergraduate theses. This can make the work stand out and leave a strong impression on professionals. By including AI technology in research, you can show flexibility and a forwardthinking approach, along with a commitment to innovation and quality. Riza Laras Amyatun and Adhan Kholis (2023) note that the impact of Artificial Intelligence (AI) on learning English is increasing. AI helps students to build English sentences. The study found that using the QuillBot AI in writing classes significantly improved students' writing skills, as shown by their test scores. The use of AI in thesis writing for BSTM students at the Glan Institute of Technology has brought different reactions from educators, parents, and students. While there are valid concerns about reduced human interaction and privacy issues, these need careful thought. However, it’s important to acknowledge the advantages of AI in improving student performance. AI can track From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5391 student progress in real time and give personalized feedback. This helps teachers see each student's strengths and weaknesses, allowing them to adjust their research work accordingly. Theoretical Stance The research is faithful to the position of Selwyn (2021) regarding educational technology inequalities. The research adopts a critical perspective in perceiving AI as not being a value-free tool but more as technology that, when used in environments with limited resources, like the Glan Institute of Technology, has the tendency to perpetuate existing inequalities. Precarious infrastructure, economic obstacles to access to improved tools, AI contextual instability (such as "hallucinations" and citations produced by the AI), and linguistic obstacles to foreign language speakers are symptoms of such deep inequalities, not technical issues. This perspective encourages us to think of fair use of AI as requiring deep institutional transformations that solve actual issues of access and enhance digital literacy, rather than simply distributing the tools. Philosophical Stance This research took a critical constructivist method. It refuses to accept that technology is impartial and sees AI as a political and social tool that works to increase disparities. This is how we contextualize problems, like infrastructure bottlenecks, spasmodic electricity, costly equipment, and discriminatory training data. These are interpreted as indications of systemic injustice and not technical errors. Dafoe (2025) concurred with this perspective by contending that the exclusions are a product of design preferences that benefit particular groups. This necessitates institutions to act in order to provide even access. This research uses this perspective by investigating how AI comes to be adopted in institutions of limited resources like the Glan Institute of Technology. The adoption is political and necessitates a conscious effort to share power of technology. Ontological Assumptions The research in this article is based on the relational theory of reality. That is, reality is constructed through continuous interaction between students lived realities and abstract outputs of AI. This is underlined by Benjamin (2022) through the demonstration of how language models based on Western cultures and cultural bias are the influences that neglect marginalized learners' realities. This research develops this analysis further by illustrating how Filipinos' indigenous knowledge becomes distorted in being translated by culturally insensate algorithms. This locates AI not as an untroubled vehicle, but as an agent in the reconfiguring of knowledge hierarchies. Epistemological Assumptions Working from a critical realist perspective, the research blames AI's hubris to construct objective knowledge by showing mechanisms under which algorithmic defect and mythical citation undermine its scholarly legitimacy. Pasquale (2023) corroborates this argument by suggesting that AI is deficient in contextual awareness to underpin technical scholarly work, convert figments of efficiency into procedures hard to check. This points to the need for human abilities, which this research considers a necessity. Students have to critically evaluate every AI response so as not to make mistakes in their perception. This renders thesis-making an intellectual vigilance all through life, rather than a fatal buildup. Axiological Assumptions This study is equityand ethics-aware and recognizes such values to be paramount in the deployment of responsible AI. Jobin (2023) refers to the care ethics in locating where algorithmic bias and linguistic issues might impact knowledge fairness. This value commitment is evident in this paper's advocacy for education interventions, namely critical AI literacy education. This exercise assists learners in validating biases in AI responses, rewording complicated vocabularies, and refuting the assumption that technology dictates results. These processes play pivotal functions in guaranteeing that AI does not impart academic exclusion. Methodological Assumptions This study utilized phenomenology to bring out lived experiences. This stands in contrast to real living conditions of students and idealistic stories of AI effectiveness. Van Manen (2024) is a believer in such a method, stating that it brings to light the "friction points" between technological promises and real struggles of people. This research employs this methodology to draw attention to underappreciated problems, for example, how power outages interfere with AI-reliant tasks and how shoestring budgets force students to seek out labor-intensive alternatives. It resets technology assessment on the basis of practical human concerns rather than theoretical objectives. Rhetorical Assumptions This research employs narrative authority to disentangle techno-solutionist discourse, successfully foregrounding students' narratives to demonstrate how AI erases certain groups. Warschauer (2023) articulated how linguistic injustice creates digital disempowerment. This research extends this theory through a comparison of institutions' demands for academic English and the multifaceted language of AI commodities. It argues that more algorithmically sophisticated technology and less communicatively explicit technology is still causing knowledge harm. This current research reinforces that technological progress should be evaluated based on how well it supports learners, rather than vice versa. From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5392 DOMAIN OF INQUIRY: One of the main challenges in incorporating AI into the creation of undergraduate thesis among 3rd Year BSTM college students at the Glan Institute of Technology is the lack of a direct connection between the subject and curriculum that has been approved. Although AI is used to improve the quality of students’ work, it is not considered a fundamental element of educational resources used in the classroom. This study aimed to address the following three key questions: 1.) How do students characterize their experiences with AI in the thesis creation process? 2.) What obstacles do students encounter when integrating AI resources into their thesis? 3.) What moral predicaments emerge for students when they use AI for thesis development? METHODOLOGY This study employed a qualitative phenomenological approach to explore the experiences of third-year BSTM students at the Glan Institute of Technology regarding the use of artificial intelligence in their undergraduate studies. To gain a thorough understanding of the participants' experiences, interviews lasting approximately two hours were conducted. A descriptive method was used to provide a detailed overview of the techniques and procedures involved, encompassing the processes of recording, transcribing, analyzing, and interpreting data. The primary methods for gathering information included in-depth interviews and focus group discussions. ETHICAL CONSIDERATIONS Prior to the informant’s participation, all third-year BSTM students provided written informed consent, which outlined their voluntary involvement, the purposes of data usage, and their unconditional right to withdraw without academic penalty. Confidentiality was ensured through pseudonymization of the interview data and secure encryption of the recordings. To keep the analysis reliable and reduce researcher bias during phenomenological interpretation, the qualitative data were subjected to structured coding. We used Creswell and Poth’s three-phase framework: open coding identified raw themes, axial coding created relational categories, and selective coding combined the main issues of algorithmic inequity and epistemic vulnerability. This organized method ensured that the participants’ experiences were grounded in knowledge and helped avoid misrepresentation. RESULTS AND DISCUSSIONS Understanding the role of artificial intelligence (AI) in education is important, because it changes the future of learning. AI technologies help personalize learning by adjusting the content to fit students' needs, which improves their outcomes. Knowing AI is becoming necessary, and understanding its strengths, weaknesses, and ethical issues is vital in a world driven by AI. Artificial Intelligence (AI) in thesis creation process AI assists college students in developing their theses by expediting literature reviews through advanced search and summarization tools. AI writing assistants refine grammar, enhance academic style, and ensure consistency. They also help organize ideas and analyze research data. However, AI remains a supportive tool and cannot replace students' critical thinking and subject expertise. IN: Unsa may imuhang naexperience the moment nga nigamit kag Artificial Intelligence in developing the concept of your paper? (What was your experience when you used Artificial Intelligence in developing the concept for your paper?) SS: Isip isa ka college student kay akong na appreciate gyud ang paggamit ug AI kay maka save siya ug oras like mas mapadali sa pag-review sa literature, pag-analize sa mga ideya na related sa thesis. Tapos mapadali ra ang proseso sa paghimo ug thesis kay dli na kailangan mag ukab-ukab pa sa libro kay ginahatag naman sa AI ang mga related study sa thesis. (As a college student, I really appreciate using AI because it saves time, like making literature review and analyzing thesis-related ideas faster. It also speeds up the thesis-writing process since I no longer need to flip through books manually because AI already provides related studies for the thesis.) IN: Aside sa napadali ang paghimog thesis, unsa pamay laing naexpeirence nimo? (Aside from making the thesis work easier, what other experiences did you have?) From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5393 SS: Nindot jud kay siya Sir ang problema lang pud ug way kuryente kay wa kay trabaho mabuhat. (It was great, Sir, but the only problem was when there was no electricity because you couldn’t do any work.) Students' experiences with artificial intelligence (AI) in thesis development show a mix of improved academic efficiency and reliance on a critical infrastructure. AI greatly speeds up literature reviews and analytical tasks by eliminating the need for time-consuming manual research. This allows students to finish their work faster, as reflected in their accounts of time saving and support for new concepts. However, the benefits provided by technology depend on reliable infrastructure. Students reported interruptions in their progress during power outages, which makes AI tools unusable because of unreliable electricity. Knight et al. (2020) highlight that while AI opens access to better research tools, its advantages depend on various factors, such as infrastructure, which could worsen inequalities in schools with fewer resources. ML: As a college student who have a thesis the role of AI is very helpful in developing our thesis. Ginapadali sa any AI sites ang pag pangita sa imong needed nga sources sa imohang thesis. Also dili AI ang ginagamit sa mga students aron mahimong mupuli sa ilahang dapat buhatunon, in fact ang AI ang ga suporta sa mga answers or ginapangita nimo sa imohang study. Maka sure man sad ta nga naay accuracy ang AI kay kini usa ka advance technology. (As a college student with a thesis, AI plays a very helpful role in developing our work. AI platforms simplify finding sources needed for the thesis. Students don’t use AI to replace their responsibilities; instead, it supports answers or information we seek for our studies. We can also trust its accuracy since it’s advanced technology.) IN: You have made mentioned accuracy because AI to you is considered advanced technology, wala ba kay kakulian o problemang nasugat.an pahanon nga nigamit kag Artificial Intelligence? (You’ve mentioned that AI is known for its accuracy because it’s advanced technology. Did you encounter any difficulties or problems when using Artificial Intelligence?) ML Naa sir, kuang kanang naay ubang AI di maaccess kay naay bayad. Kanang abi nimog free. Unya excited ka sir kay imuhang gipangita mao answer sa imuhang problema sa kaso di nimo macopy ang tibuok nga answer kay need nimog subscription, unya mahal raba kaayo. (Yes, Sir. One issue is that some AI platforms aren’t accessible because they require payment. You think they’re free, so you get excited when you find the answer to your problem. But then you realize you can’t copy the full response because a subscription is needed and it’s expensive.) IN: Unya kung ingun ana ang panghitabo, unsay ginabuhat nimo? (So when that happens, what do you do?) ML: Mangita napug laing sites nga naay free sir unya, magbasa napud ug answers nga gihatag ug mao bay tubag sa pangutana nga kinahanglan nako maansweran. (When that happens, Sir, I look for other websites that offer free access. I read the answers they provide and check whether they match the question I need to answer.) ML's interaction with paywalled AI tools, described as "naay ubang AI di maaccess kay naay bayad," highlights an important financial aspect of exclusion within academia. Economic barriers worsen inventors ‘structural weaknesses. The frustration of facing subscription requirements after putting in mental effort, referred to as " or di nimo macopy," shows what Van Deursen and Van Dijk (2019) call material access inequality. This shift moved the digital divide from physical connection issues to financial support gaps. As ML turns to inefficient workarounds, such as looking for other free platforms, "mangita napug laing mga sites,” it demonstrates how commercial AI tools unintentionally create tiered access systems. These systems force students with fewer resources into disconnected research processes, which reduces the time savings offered by AI. As a result, ML's experiences with AI have expanded the electricitydependency narrative into a broader framework. This reveals how infrastructure problems and financial barriers divide the educational benefits of AI globally. From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5394 Obstacles in integrating Artificial Intelligence (AI) resources Integrating AI into education poses challenges for under resourced areas. This worsens the digital divide and limits access to hardware, the Internet, and devices. The shortage of technical support and digital literacy training for students makes effective adoption even more challenging. Many students in these resource-limited settings lack the skills required to use AI systems. IN: In developing the concepts of your thesis and the different chapters, unsa may pinakalisud nga imuhang naagian panahon nga nagtrabaho mo sa inyohang papel? (What was the most difficult experience you had while working on your thesis and its various chapters?) HE: Oras sir, labi nag muhatag kag deadline, kay deadline is deadline man jud ka sir human usahay labad pud ang tubag sa AI sir. (Time, sir, especially when you set a deadline, because you're strict with deadlines and sometimes the AI gives confusing answers.) IN: Unsay bout pasabot nimo anang labad ang tubag usahay ang tubag sa AI? (What do you mean by the AI giving confusing answers?) HE: Muhaluucinate siya sir ay, Kanang gusto kag quick response aron mahuman naka sa imuhang gitrabaho, di niya ihatag ang tamang answer. (It hallucinates, sir. When you want a quick response to finish your work, it doesn't give the correct answer.) IN: Hatagi kuno kog isa ka specific example (Can you give me one specific example?) HE: Pareho atong nagdevelop mig concept about tourism ug relationship sa environment, ang iyahang response lagyo kayo sir, dili specific. (When we were working on the concept of tourism and its connection to the environment, AI's response was completely off and not detailed.) The transcript excerpt shows HE's struggles with the AI-generated "hallucinations" during thesis development. Specifically, HE received irrelevant responses when asked about tourism-environment relationships. This highlights significant reliability issues in the use of AI in education. This echoes Shneiderman's (2020) claim that AI systems that do not follow strong human-centered design principles often fail to provide suitable outputs, especially for complex academic tasks. Students feel frustrated by vague answers under time pressure. This reveals how poorly designed tools can reduce efficiency even when they claim to save time. Artificial intelligence does not involve emotions, subjective understanding, or contextual sensitivity. It operates only using data-driven algorithms and pattern recognition. This approach can lead to outputs that do not meet nuanced academic needs or user intentions, as shown in the transcript below. GG: Usa ka challenges sa pag gamit ug AI sa thesis kay ang lisod nga pag pangita sa RRL kay dili sad tanan naa sa AI or ma search sa AI kay related sa imohang topic sa thesis so mag lisod gyud ka ug pinangita. One challenge in using AI for this thesis is finding a Review of Related Literature (RRL). Not all sources are available through AI or searchable by AI, especially those closely related to your thesis. This makes the search more difficult. IN: Pano man ninyo mahibaw.an nga dili mao ang reference nga gihatag sa AI? (How do you know if the reference provided by AI is incorrect?) GG: Kuan sir katong nagklase ta one time, naremember namo nga naka ingun ka nga pag online ang paper, kanang napublish siya, naay DOI sir. (Well, Sir, during our class one time, we remembered you mentioned that for online papers, if it's published, it has a (DOI) Digital Object Identifier). IN: Unya unsay naa sa DOI? (And what about the DOI?) GG: Kung sir kanang inig click nimo, di malocate sir, wala siyay website, meaning ana sir hinimo-himo lang sa AI sir, dili tinood ang reference. (Sir, when you click it, the website can't be located meaning it has no valid website that indicates the AI fabricated it, Sir. The reference isn't real.) From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5395 GG's experience of encountering fabricated references "hinimo-himo lang sa AI" while searching for a Review of Related Literature (RRL) highlights a significant integrity gap in AI-assisted academic research. This issue arises when tools generate nonexistent sources owing to limitations in the training data or algorithmic hallucinations. This observation aligns with Schiff's (2022) analysis of epistemic unreliability in educational AI, where systems lacking domain-specific validation mechanisms produce outputs that are disconnected from scholarly authenticity. The student's use of DOI verification "ig click nimo, di malocate... dili tinood", as taught in class, illustrates how resource-constrained learners must develop ad hoc critical literacy skills to compensate for AI's shortcomings, paradoxically increasing their cognitive labor despite AI's promise of efficiency. IN: In using artificial intelligence (AI), unsa may imuhang pinakalisud nga problemang naagian? (In using artificial intelligence (AI), what has been the most difficult problem you’ve encountered?) XX: naglisod ko pag sabot sa grammar sa AI sir. (I struggle to understand the grammar used by the AI, sir.) IN: Nganu man, nga English man ang tubag as AI. (Why is that, when the AI responds in English?) XX: dili nako first language ang English sir, sa AI mostly gagamit siyag complex term or vocabularies sa result sa amoang gina pang research maong usahay mag lisod ta ug sabot. Bulok bulok pud bya ta sir, hinay kay ta anang vocabulary maong kung unsay ihatag sa AI need jud basahon sir kay ang answer lag-gom kaayog English. (English isn’t my first language, sir. The AI often uses complex terms or vocabulary in the results of our research, which makes it hard to understand. We’re not that fluent, sir our vocabulary is weak, so whatever the AI gives, we really must read it thoroughly because the answer is written in very deep English.) IN: Dili diay simple sentences? (So, it doesn’t use simple sentences?) XX: Dili sir oi, ug muhatag syag answer kinahanglan jud nimo basahon kay di man siya mao jud dayon, usahay mura kag robot ug basahon nimo way feeling. (No sir it doesn’t. When it gives an answer, you really must read it carefully because it’s not immediately clear. Sometimes it feels like you’re reading like a robot there’s no emotion in it.) The challenges faced by XX in interpreting AI-generated outputs characterized by complex vocabulary "gagamit siyag complex term or vocabularies" and robotic phrasing "mura kag robot ug basahon... way feeling", compounded by the fact that English is not their first language "dili nako first language ang English", highlight a significant linguistic equity gap in AI educational tools. This issue aligns with Kohnke (2023) findings, which indicate that generative AI frequently defaults to lexically dense, native-level English that is inaccessible to ESL learners, thereby creating a "lexical barrier" that necessitates exhaustive decoding "kinahanglan jud nimo basahon" and increases cognitive load. The disparity between AI’s Western-trained language models and the proficiency levels of learners from the Global South "bulok bulok pud bya ta" exemplifies how ethnolinguistic biases intensify educational inequity. Moral Predicaments on AI use in thesis development The adoption of AI in thesis writing creates ethical tensions around authenticity and academic integrity, as students risk presenting AI-generated content, such as fabricated references or misaligned outputs, as their original work. This blurs authorship boundaries and undermines accountability when time pressures prioritize speed over verification. Overreliance on AI tools may erode critical thinking by limiting research engagement. JB: Isa ka negative effects sa quality sa research when using AI is the originality, ang pag gamit ug sobra sa AI kay makaresulta ug tampered or di na original nga idea kay puro naman ka ga rely sa AI. (One negative effect on research quality when using AI is the loss of originality; using AI too much can result in tampered or non-original ideas because you're just relying on it entirely.) IN: In what way man nimo na notice ang ingun ani nga panghitabo? (In what way did you notice this kind of situation?) From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5396 JB’s observation that excessive reliance on AI diminishes originality "puro naman ka ga rely sa AI... tampered or di na original nga idea" and encourages superficial engagement "maghangad hangagd ra kay way nasabtan" underscores a significant ethical dilemma between efficiency and genuine scholarship. Students run the risk of swapping their genuine intellectual work for the content created by AI. This aligns with Krügel et al. (2023) finding that AI tools can lead to "effort substitution," where learners bypass deep cognitive engagement by uncritically accepting AI content, resulting in fragile comprehension and compromised academic integrity. The tendency of JB’s peers to copy-paste from ChatGPT "basta naka pasa, basta copy paste gikan chat gpt” illustrates how AI’s convenience promotes passive plagiarism, especially under deadline pressure. This situation calls for a pedagogical redesign that requires a critical evaluation of AI outputs and redefines assessment to focus on process transparency rather than mere product expediency. IN: Unsaon sa mga estudyante pagpanalipod sa ilang research findings batok sa posibleng bias nga gikan sa paggamit sa AI algorithms? (How can students protect their research findings against potential bias from using AI algorithms? LA: Ang AI algorithms mahimong adunay biases nga makaapekto sa resulta sa research. Ang mga estudyante kinahanglan nga magbantay ug mag-review sa mga resulta aron masiguro nga wala’y biases nga naghatag ug dili makatarunganon nga epekto sa ilang research findings. (AI algorithms may contain biases that can affect research outcomes. Students should vigilantly review results to ensure no biases create unfair impacts on their findings.) IN: Hatagi kuno kog example? (Could you give an example?) LA: Kuan sir, ang AI kay pwede mamanipulate ang answer, depende ug unsaon nimo pag promp niya, ug gusto kag positive explanation sa data muhatag siya ug gusto nimo negative ang interpretation okay ra sab siya. (Well, sir, AI can manipulate answers depending on how you prompt it. If you want a positive explanation of the data, it provides one; if you prefer a negative interpretation, that’s also possible.) IN: So in what way nimo siya ginagamit properly? (So how do you use it properly?) LA: Akoa jung ginatandaan imuhang tudlo sir nga magbasa, magpaphrase unya tanggalon ang mga wala nay apil sa discussion dapat. Pero infairness sir kanang magbuhat kag explanation dali nalang jud pag naay AI dali ra makahimo kay naay guide. (I follow your guidance, sir: read, paraphrase, then remove irrelevant parts from the discussion. But in fairness, sir, crafting explanations becomes much easier with AI it’s quick to generate drafts when you have a guide.) JB: Kuan sir, kanang kung basahon nimo ang response niya, nindot man siya pero nawala ang human touch, wala siyay feeling, maong need jud sya ipaphrase para mualign siya sa insaktong tubag. (Well, sir, when you read its response, it's well-written but lacks the human touch there’s no emotion, so it needs to be rephrased to align with the proper answer.) IN: Does this affect originality as you have mentioned ganiha? (Does this affect originality as you mentioned earlier?) JB: Yes sir oi, pareho atong nagpareporting ka sir sa amoa panahon nga nag develop Patag concept, akoang mga ubang member sa group maghangad hangagd ra kay way nasabtan, kay basta lagi naka pasar a basta copy paste gikan chat gpt. (Yes sir, like when you asked us to do a report while we were developing the concept. Some of my group members just stared blankly because they didn’t understand anything. They just submitted it, even if it was copy-pasted from ChatGPT.) From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5397 LA’s recognition that AI algorithms can introduce biases into research outcomes "mahimong adunay biases nga makaapekto sa resulta" and require critical evaluation "magbantay ug mag-review sa mga resulta" highlights the ethical necessity of maintaining algorithmic vigilance in AI-assisted scholarship. This aligns with Zawacki-Richter et al. (2019) warning that AI systems tend to have latent biases from training data or prompt engineering that, if not controlled, will undermine research objectivity. LA's mitigation strategy, "magbasa, magpaphrase, unya tanggalon ang wala nay apil", suggests that learners can use critical AI literacy to take over bias, turning AI into a guided drafting tool "dali ra makahimo kay naay guide". But this task is outsourcing the role of a spotting bias to resource-deprived learners. Institutions must also develop concise guidelines for AI design document and ensure proper training in prompt auditing practices. CONCLUSIONS This qualitative study shows that while AI applications can immensely improve thesis composition for BSTM students by speeding up literature reviews, making composition more accurate, and data analysis simpler, their use is riddled with issues. The benefit of AI tools depends on availability of reliable infrastructure and universal access to cheap inputs. The digital divide and economic polarization can turn AI into an instrument of inequality in health care and not an instrument of support. This study clearly manifests what the major shortcomings of current AI systems, such as incorrect citations and context-inappropriate responses are. These reliability problems cause students to double-check information very unnecessarily, generally eliminating time gained, and to make new skill to assess AI-generated work. The paper concludes that the pedagogical and moral issues caused by AI in thesis composition necessitate institution and teaching levels of active interventions. Overdependence on AI threatens critical thought, individuality, and scholarship, as has been unfolding in the directions towards shallow approach and effort replacement. Biases within AI algorithms need students' keen examination to safeguard their integrity. Instructional practices must be adapted in ways that responsibly utilize AI. It entails creating critical AI literacy, altering assessment modes to prioritize process transparency and engagement, and pushing for institutional support to bridge access gaps. AI must be utilized as a supportive facility in navigating the thesis process but not at the expense of critical thinking, ethical reasoning, and genuine scholarly inputs. IMPLICATIONS FOR EDUCATIONAL PRACTICE Incorporating AI into thesis writing involves fundamental shift in pedagogical practices to optimize its potential while minimizing risks. Schools must develop curricula that educate students on critical AI literacy. This will enable students to become proficient in how to assess AI results for truthfulness, bias, and ethical purposes. Students must authenticate their sources and utilize tools as assists, not substitutes for their intelligence. Testing procedures should be reworked to focus on the research process. This entails checking sources, comparing information, and coming up with various drafts, rather than just evaluating the final product. This dissuades shallow use of AI-generated content. Along with it, educational policies should bridge the infrastructure, and equity divides by providing students secure access to technology and affordable AI tools. It is also a requirement that faculty members are trained so that teachers can assist students in learning the limitations of AI and academic integrity. Lastly, AI must be kept as an auxiliary tool in instruction. This ensures that the paramount functions of human critical thinking, ethical consideration, and authentic scholarly contribution are not compromised. REFERENCES 1) Alharbi, W. (2023). AI in the Foreign Language Classroom: A Pedagogical Overview of Automated Writing Assistance Tools. Education Research International, 2023, 1–15. https://doi.org/10.1155/2023/4253331 2) Benjamin, R. (2022). Decolonial AI: Unmasking algorithmic violence in education. AI & Society, *38*(3), 1229– 1241. https://doi.org/10.1007/s00146-022-01548-2 3) Creswell, J. W., & Poth, C. N. (2018). Qualitative inquiry and research design: Choosing among five approaches (4th ed.). SAGE. 4) Dafoe, A. (2025). The embedded politics of AI: Power, justice, and technological design. Science, Technology, & Human Values, *50*(1), 78–102. https://doi.org/10.1177/01622439241234567 5) Jobin, A. (2023). Beyond fairness: Toward an ethics of care for educational AI. Learning, Media and Technology, *48*(4), 632–647. https://doi.org/10.1080/17439884.2023.2234567 6) Knight, S. R., Skouras, S., & Dalton, R. (2020). The role of artificial intelligence in enhancing clinical research efficiency. Nature Digital Medicine, 3(1), 57. https://doi.org/10.1038/s41746-020-0260-4 From Efficiency to Ethics: A Phenomenological Inquiry Into AI-Enhanced Thesis Writing Among BSTM Students IJMRA, Volume 08 Issue 09 September 2025 www.ijmra.in Page 5398 7) Kohnke, L. (2023). L2 learners’ perceptions of using ChatGPT for writing support: Affordances and challenges. RELC Journal. Advance online publication. https://doi.org/10.1177/003368822311628 8) Krügel, S., Ostermaier, A., & Uhl, M. (2023). ChatGPT’s impact on students’ critical thinking and learning motivation. Journal of Computer Assisted Learning, 39(6), 1902–1916. https://doi.org/10.1111/jcal.12887 9) Nazari, N., Shabbir, M. S., & Setiawan, R. (2021). Application of Artificial Intelligence powered digital writing assistant in higher education: randomized controlled trial. Heliyon, 7(5), e07014. https://doi.org/10.1016/j.heliyon.2021.e07014 10) Pasquale, F. (2023). Data influenced knowledge: The limitations of AI in specialized domains. Social Epistemology, *37*(4), 456–472. https://doi.org/10.1080/02691728.2023.2193456 11) Riza Laras Amyatun, & Adhan Kholis. (2023). Can Artificial Intelligence (AI) like QuillBot AI Assist Students’ Writing Skills? Assisting Learning to Write Texts using AI. English Language Education Reviews, 3(2), 135–154. https://doi.org/10.22515/elereviews.v3i2.7533 12) Selwyn, N. (2021). Ed-Tech within limits: Anticipating educational technology in times of environmental crisis. Learning, Media and Technology, 46(3), 364–377. https://doi.org/10.1080/17439884.2021.1942793 13) Schiff, D. (2022). Education in the age of artificial intelligence: Promise and pitfalls. Routledge. https://doi.org/10.4324/9781003282002 14) Shneiderman, B. (2020). Human-centered artificial intelligence: Reliable, safe & trustworthy. International Journal of Human–Computer Interaction, 36(6), 495–504. https://doi.org/10.1080/10447318.2020.1741118 15) Van Deursen, A. J. A. M., & Van Dijk, J. A. G. M. (2019). The first-level digital divide shifts from inequalities in physical access to inequalities in material access. New Media & Society, 21(2), 354–375. https://doi.org/10.1177/1461444818797082 16) Van Manen, M. (2024). Doing phenomenological research in the age of AI: Human experience beyond algorithms. Qualitative Inquiry, *30*(3), 345–359. https://doi.org/10.1177/10778004241234567 17) Warschauer, M. (2023). AI and the digital language divide in global education. Language Learning & Technology, *27*(2), 1–18. https://doi.org/10.5281/zenodo.7890123 18) Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review of research on artificial intelligence applications in higher education: Where are the educators? International Journal of Educational Technology in Higher Education, 16(1), 1–27. https://doi.org/10.1186/s41239-019-0171-0 There is an Open Access article, distributed under the term of the Creative Commons Attribution – Non Commercial 4.0 International (CC BY-NC 4.0) (https://creativecommons.org/licenses/by-nc/4.0/), which permits remixing, adapting and building upon the work for non-commercial use, provided the original work is properly cited.