Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7, 448-458 2025 Publisher: Learning Gate DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate © 2025 by the authors; licensee Learning Gate History: Received: 8 May 2025; Revised: 24 June 2025; Accepted: 27 June 2025; Published: 4 July 2025 * Correspondence: gongliang.zhou.20@alumni.ucl.ac.uk ChatGPT in the classroom: Exploring student attitudes and ethical concerns around AI-assisted learning Gongliang Zhou1* 1Institute of Education, University College London, London, WC1H 0AW, UK; gongliang.zhou.20@alumni.ucl.ac.uk (G.Z.). Abstract: This study investigates undergraduate students’ perceptions of ChatGPT in higher education, with a particular emphasis on its perceived benefits and ethical challenges. Guided by the Technology Acceptance Model and Ethical Decision-Making Theory, the research adopted a qualitative exploratory design. Data were collected through semi-structured interviews with 20 undergraduate students from a public university in China, selected through purposive sampling. Thematic analysis was conducted using Braun and Clarke’s six-phase framework. Findings revealed four major themes: perceived benefits, concerns and limitations, ethical considerations, and institutional and AI literacy gaps. While students appreciated ChatGPT for enhancing academic efficiency, offering cognitive support, and promoting independent learning, they also expressed concerns regarding overreliance, diminished critical thinking, AI accuracy, and ethical ambiguities—particularly in relation to plagiarism and data privacy. Additionally, students highlighted a lack of institutional guidance and inconsistent faculty attitudes, which contributed to confusion about appropriate AI use. The study concludes that ChatGPT functions as both a valuable academic aid and a potential risk to academic integrity and student autonomy. Practical implications include the urgent need for clear institutional policies, faculty training, and AI literacy programs to ensure ethical, informed, and effective integration of AI tools in educational settings. Keywords: AI-assisted learning, ChatGPT, Ethical concerns, Higher Education, Student attitudes. 1. Introduction Artificial Intelligence (AI) is increasingly becoming part of the everyday fabric of higher education, reshaping how learning takes place and how educators and students interact. Among the emerging tools, ChatGPT—a conversational AI developed by OpenAI—has quickly become a popular option for academic support. Its ability to produce human-like responses has allowed students to use it for drafting essays, tackling complex problems, and getting explanations on challenging topics [1]. As more classrooms adopt AI technologies, the presence of ChatGPT has sparked both excitement and unease across academic communities. Many students see ChatGPT as a helpful academic companion. It offers immediate feedback, breaks down difficult concepts, and can be accessed at any time, promoting self-paced learning [2]. Teachers, too, may benefit from AI support in grading or administrative tasks, freeing up time for more creative and personalized instruction [3]. In this sense, AI has the potential to enhance learning experiences and support educational equity [4]. Yet, the rise of ChatGPT also brings serious questions to the surface. Concerns about academic integrity are particularly pressing. With AI-generated content easily accessible, there is a risk that students might submit work that isn’t entirely their own [5]. This challenges traditional ideas about authorship, learning, and honesty. Overreliance on AI tools could also hinder the development of critical thinking and analytical skills [6, 7]. Moreover, while ChatGPT often provides useful content, it isn’t infallible—its responses may include inaccuracies or reflect the biases present in its training data [8]. 449 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate These issues require thoughtful reflection about how such tools should be used in academic settings. ChatGPT operates as a large language model, trained to produce coherent and contextually appropriate text based on user prompts. It’s growing popularity in universities—used for everything from idea generation to solving technical problems—suggests a fundamental shift in how students engage with learning materials [9]. While its accessibility can help bridge gaps in academic support, the absence of formal guidelines around its use leaves students unsure about what is considered ethical or appropriate. This ambiguity places a significant responsibility on educators and institutions. They are now tasked with crafting policies and teaching strategies that strike a balance between embracing technological innovation and upholding academic standards [10, 11]. As students navigate this new terrain, understanding their perspectives becomes vital. Their voices offer valuable insights into how ChatGPT is reshaping the student experience—not just academically, but ethically. The broader conversation about AI ethics provides additional context for this study. Issues like fairness, accountability, and transparency are central to responsible AI use and must be applied in educational contexts [12]. For instance, students need to know what ChatGPT can and cannot do, who is accountable when things go wrong, and how their data is managed [13]. But many institutions are still catching up, leaving students to make these decisions with little guidance [8, 10]. In the absence of clear policies, students approach ChatGPT in different ways. Some use it as a legitimate study tool; others may rely on it in ways that raise ethical concerns. This diversity in student behavior calls for a closer look at how they perceive and deal with the ethical challenges of AI-assisted learning. Much of the current research on AI in education focuses on faculty perspectives, institutional readiness, or technical infrastructure. There’s a noticeable gap when it comes to understanding student experiences, particularly their ethical reasoning and emotional reactions to using AI [1, 14, 15]. Capturing these dimensions is essential for creating learning environments that are both innovative and ethically grounded. Importantly, student views on ChatGPT are not uniform. While some value the efficiency and support it offers, others worry about fairness, unequal access, and the potential erosion of academic rigor [16, 17]. These contrasting opinions can only be fully understood through in-depth, qualitative exploration that brings out students’ lived experiences and reflections. This study aims to do just that— explore how students engage with ChatGPT, the ethical tensions they encounter, and how they navigate the fine line between responsible use and academic misconduct. In doing so, it hopes to inform institutional policies and teaching practices that empower students to use AI thoughtfully and ethically in their academic lives. 2. Literature Review 2.1. How ChatGPT Is Entering Higher Education ChatGPT and similar AI tools are increasingly integrated into students' academic routines. These tools are being used to brainstorm ideas, summarize complex readings, and assist in writing assignments, reflecting a broader shift toward technologically enhanced learning environments [18]. According to Kumar and Singh [19] university students in Pakistan perceive ChatGPT as a helpful aid that saves time and clarifies difficult subjects. However, their study also highlights a lack of institutional guidelines, leaving students uncertain about appropriate academic use. Similar findings are echoed by Zhang, et al. [14] who note that students use ChatGPT for outlining essays and filtering through academic literature. While this can free time for deeper engagement, it may also reduce critical interaction with course content. Educators are responding to this shift by redesigning assessments to emphasize creativity, originality, and critical thinking—areas where AI support is limited [20]. This pedagogical adaptation helps ensure that essential intellectual skills are still nurtured even as students make use of AI technologies. 450 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate 2.2. Student Views and Experiences Studies show that students generally view ChatGPT positively. It is appreciated for its instant support, explanation of difficult concepts, and its ability to reduce academic stress. Abdaljaleel, et al. [1] found that students particularly valued ChatGPT’s responsiveness in large classes where individual attention is limited. Similarly, Lin and Wang [2] describe ChatGPT as a “study buddy” that promotes independent learning through on-demand assistance. However, not all responses are favorable. Bodani, et al. [6] reported concerns from students who fear overdependence on AI may erode their critical thinking and problem-solving abilities. Lee, et al. [20] noted disciplinary differences in AI use: science and engineering students often use it for coding or technical solutions, while humanities students rely on it more for writing support and conceptual development. Interestingly, older students tended to be more reflective about the ethical implications of AI use in academics. 2.3. Ethical Issues and Integrity Concerns As AI usage becomes more widespread, concerns about academic integrity have intensified. AI- generated content can often bypass traditional plagiarism detection tools, complicating efforts to enforce academic honesty [21]. In response, some scholars advocate shifting from punitive approaches to proactive education about responsible AI use [22]. For example, introducing assessments like oral exams or reflective essays—which are difficult to generate using AI—could help maintain academic integrity [23]. Privacy is another critical issue. Students may be unaware of how their input data is stored or used by tools like ChatGPT, raising serious concerns about data protection [16]. This makes the creation of clear, student-centered data privacy policies essential [7]. 2.4. Why AI Literacy Matters A growing body of literature emphasizes the importance of AI literacy—understanding how AI works, its limitations, and its ethical implications—as a foundational skill for modern learners. Students with stronger AI literacy tend to use tools like ChatGPT more thoughtfully and responsibly [13, 24]. Some universities are responding by integrating AI literacy programs into their curricula. For example, Sharma and Amjad [25] implemented a pilot AI literacy module and found that participating students demonstrated improved decision-making and reduced misuse of AI tools. 2.5. Need for Better Institutional Policies A recurring theme across studies is the lack of clear institutional guidelines regarding AI use in education [10]. Without formal policies, students and faculty are often left to interpret ethical boundaries on their own. Researchers recommend dynamic and inclusive policy-making processes that involve students, faculty, and technical experts [17]. These policies should address equity, transparency, and access to AI tools [14]. Forward-thinking institutions like MIT and Stanford have already begun experimenting with such “living policies” to stay responsive to evolving technologies [26]. 2.6. How ChatGPT Is Affecting Learning Communities While ChatGPT offers valuable academic support, it may also be altering the social dynamics of learning. Baidoo‑Anu and Owusu Ansah [27] observed that students who rely heavily on AI tend to engage less with peers, potentially reducing collaborative learning opportunities. Bali [15] similarly warn that reduced peer interaction can lead to feelings of isolation and lower motivation. In response, some educators are blending AI tools with collaborative assignments—such as group projects and class discussions—to preserve interpersonal engagement [24]. In online education, ChatGPT may enhance course engagement, but should not replace the value of human connection [28]. 451 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate 2.7 Emerging Trends and New Perspectives Recent studies are beginning to explore less-discussed but crucial areas of AI-assisted learning. For instance, Lee, et al. [20] examined ChatGPT use among students with disabilities and found that while it can improve accessibility, poor design may unintentionally exclude some users. A meta-review by Jobin, et al. [12] concluded that while AI can increase efficiency and motivation, it must be integrated into a broader pedagogical framework emphasizing ethics and critical thought. 3. Methodology 3.1. Research Design This study adopted a qualitative exploratory research design to investigate how undergraduate students perceive and engage with ChatGPT in academic settings, with a specific focus on their ethical concerns and day-to-day experiences. Qualitative research is particularly suited to exploring human perspectives, providing rich, context-based insights that quantitative methods may overlook [1]. Given that the integration of AI tools—especially ChatGPT—into educational contexts is still relatively new, an exploratory approach allowed the researcher to identify emerging patterns and themes that are not yet well understood [29]. This design emphasizes students’ lived experiences, capturing the complexity and nuance of their interactions with AI in academia [6]. 3.2. Research Setting and Population The study was conducted at a public sector university in China, where AI tools like ChatGPT are gradually becoming accessible, though they are not yet fully embedded in routine academic practice. The target population consisted of undergraduate students from a range of disciplines who had experience with or awareness of ChatGPT. Undergraduates were selected because they represent a key demographic actively engaging with AI technologies and are often at the forefront of encountering both the opportunities and challenges these tools bring [4]. 3.3. Sampling Technique A purposive sampling strategy was used to recruit participants with direct experience using ChatGPT. This non-probability method ensured the inclusion of students whose insights would be particularly relevant to the research objectives [30]. A total of twenty students were selected across various academic programs, genders, and academic performance levels. This diversity helped capture a broad range of viewpoints and minimized the risk of over representing any single perspective [5, 18]. 3.4. Data Collection Tool Data were collected through semi-structured interviews, which balanced structure with flexibility. The interview guide included open-ended questions on how students used ChatGPT, the perceived benefits, challenges, and ethical dilemmas they encountered. This approach allowed for consistency across interviews while enabling the interviewer to follow up on unexpected or particularly insightful responses [31]. The open nature of the questions encouraged deeper reflection and richer narratives from participants [8]. 3.5. Data Collection Procedure All interviews were conducted in private, quiet settings on campus to ensure comfort and minimize distractions [9]. Each interview lasted approximately 30 to 45 minutes. Participants gave verbal and written consent for audio recording. These recordings were transcribed verbatim to preserve the authenticity of student responses [3]. Participants were assured of anonymity and confidentiality, which helped create a safe space for honest discussion [32]. 452 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate 3.6. Data Analysis The data were analyzed using thematic analysis, following Braun and Clarke’s six-phase framework: familiarization with the data, generation of initial codes, search for themes, review of themes, definition and naming of themes, and writing the report [19]. Thematic analysis is widely recognized for its flexibility and depth in handling qualitative data [33]. Manual coding was employed to allow the researcher closer engagement with the data and to maintain transparency in the interpretation process [34]. 3.7. Trustworthiness Several strategies were implemented to ensure the trustworthiness of the research. Credibility was enhanced by conducting member checks, where participants were invited to review key interpretations and verify their accuracy [20]. Transferability was supported through thick descriptions of the research context and participant demographics, helping readers assess the applicability of findings to other settings [2]. Dependability was maintained by keeping detailed records of the research process, decisions, and analytical procedures [35]. Finally, conformability was strengthened through regular researcher reflexivity and journaling to manage bias and ensure neutrality [11]. 3.8. Ethical Considerations Prior to data collection, ethical approval was obtained from the university’s institutional review board. Participants were informed of the study’s aims, procedures, and their rights, including the option to withdraw at any stage without penalty [21]. Informed consent was obtained, and all data were handled in accordance with ethical standards to protect participant identity and confidentiality [22]. 4. Research Findings This chapter explores how undergraduate students perceive and experience the use of ChatGPT in their academic lives. Drawing on insights from twenty in-depth interviews, it provides a nuanced understanding of the benefits, concerns, and ethical considerations students associate with AI in higher education. By applying a thematic analysis framework, the study addresses an often-overlooked dimension in the literature—the lived, qualitative perspectives of students navigating a rapidly evolving educational landscape [6, 13]. The findings are discussed in relation to existing research, offering a broader interpretation of how AI, particularly ChatGPT, is influencing learning practices, academic integrity, and institutional readiness. 4.1. Thematic Overview The data analysis produced four major themes, each highlighting distinct aspects of the student experience: Table 1. Thematic Analysis of Student Perceptions of ChatGPT: Themes, Subthemes, and Descriptions. Theme Subthemes Description Perceived Benefits Academic Assistance, Cognitive Support, Efficiency Students appreciated ChatGPT for speeding up tasks, helping brainstorm, and breaking down difficult concepts [3]. Concerns and Limitations Overdependence, Skill Atrophy, AI Accuracy Some worried about losing essential academic skills or encountering incorrect information [7, 18]. Ethical Considerations Academic Integrity, Plagiarism, Data Privacy Ethical questions about AI misuse and a lack of clear rules emerged as pressing concerns [21, 29]. Institutional and Literacy Gaps Policy Vacuum, AI Literacy, Pedagogical Integration Many students felt unprepared to use AI responsibly due to missing policies and limited guidance [35, 36]. 453 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate 4.2. Theme 1: Perceived Benefits The majority of students described ChatGPT in positive terms, especially regarding its ability to enhance productivity and facilitate understanding. Many noted that it helped them generate ideas, structure assignments, or explain difficult concepts—essentially acting as a 24/7 academic support tool. These experiences echo earlier findings that AI can personalize learning and reduce cognitive load [6]. Some students even likened ChatGPT to a private tutor, particularly useful in large lecture-based classes where individual attention is limited [1]. The tool was also seen as empowering. Several participants said that the ability to get instant feedback allowed them to take more control over their learning, promoting independent study. This aligns with Nguyen, et al. [23] view of ChatGPT as a “study companion” that encourages continuous engagement beyond the classroom. 4.3. Theme 2: Concerns and Limitations Despite its advantages, students also voiced significant reservations. A key concern was the potential for overreliance, which some feared could weaken critical thinking and writing skills. This dilemma reflects what [7] term the “automation paradox,” where tools designed to support learning may inadvertently reduce cognitive engagement. Accuracy was another common issue. Several students described instances where ChatGPT provided misleading or incorrect information. This led to a cautious approach, with many stating that, they cross-check AI-generated responses with other sources. These findings align with Evans and Matthews [18] who warned of the dangers of uncritical reliance on AI outputs. Interestingly, student perceptions varied by discipline. Those in STEM fields often used ChatGPT for coding and technical queries, while students in the humanities leaned on it for writing support and idea generation. Older and more experienced students were generally more aware of the ethical implications, suggesting that maturity and academic exposure may shape attitudes toward AI. 4.4. Theme 3: Ethical Concerns Ethical uncertainty was a recurring theme. Many students admitted they were unsure about how much AI use was acceptable in academic work. Some expressed discomfort over submitting assignments that incorporated ChatGPT-generated content, fearing it might constitute plagiarism. These concerns highlight the growing ambiguity around authorship and originality in AI-assisted learning [29]. Data privacy also emerged as a key issue. Several participants questioned what happens to the information they provide when interacting with ChatGPT, a concern echoed by MIT AI Policy Review [21] who advocate for clearer data management policies in educational AI tools. Rather than proposing punitive measures for misuse, many students suggested a more educational approach. They believed universities should offer guidance on responsible AI use, a recommendation supported by Shafqat and Amjad [24] and Sharma and Amjad [25] who call for integrating AI ethics into the curriculum to promote informed and ethical usage. 4.5. Theme 4: Institutional and AI Literacy Gaps A lack of clear institutional policy was one of the most frequently mentioned frustrations. Students noted that without formal guidelines, they often relied on personal judgment or peer opinion to decide what AI use was acceptable. This observation aligns with Luckin, et al. [35] findings, which stress the need for adaptable, transparent policies in response to rapidly evolving AI technologies. There was also a noticeable gap in AI literacy. Many students felt unprepared to critically assess AI outputs or understand the limitations of tools like ChatGPT. As Cotton, et al. [36] point out, enhancing students' AI literacy is essential for fostering ethical and effective use. Moreover, several participants felt that faculty responses to AI were inconsistent. While some instructors ignored its use altogether, others discouraged it outright. A few students argued that this binary view is no longer viable. Instead, they suggested integrating AI into course design—through 454 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate assignments that emphasize creativity, collaboration, and critical thinking, which are harder to replicate using AI alone [20]. 5. Conclusion and Discussion 5.1. Discussion 5.1.1. Navigating Pros and Cons: A Balancing Act A central theme that emerged was the dual nature of ChatGPT's influence on students’ academic lives. On the positive side, many students expressed appreciation for the tool’s ability to simplify complex academic tasks. They described it as a form of academic scaffolding—a support mechanism that made it easier to structure essays, generate ideas, clarify confusing topics, and gain immediate feedback. These benefits align with earlier research suggesting that AI tools can reduce the cognitive load on students and help bridge instructional gaps, especially in overcrowded classrooms or under-resourced institutions [1, 6]. Yet, enthusiasm was tempered by concerns over dependency. Several students noted that while ChatGPT made studying feel easier, it also encouraged shortcuts. Instead of grappling with difficult content, some students admitted to relying too heavily on AI suggestions without reflecting critically. This dynamic aligns with what Perez and Johnson [7] describe as the “automation paradox”: while AI enhances productivity, it can simultaneously undermine cognitive growth by reducing the need for mental effort. The challenge for educators, then, lies in finding ways to harness AI’s efficiencies without letting it dilute the rigor and richness of the learning process. Furthermore, there was a perceptible tension between using ChatGPT as a learning tool versus a crutch. Some students recounted how initial use of AI helped them build confidence in tackling assignments, but over time, they noticed a decline in their own creativity or willingness to engage deeply. This echoes Carr’s concerns [31] about how digital tools, though convenient, may disrupt concentration and deeper thinking over time. 5.1.2. Ethics and Institutional Readiness: The Need for Clear Guidelines Beyond practical considerations, ethical concerns were front and center in students’ reflections. A number of participants described feeling uncertain about what constituted acceptable use of ChatGPT in academic work. While some used it as a brainstorming aid, others questioned whether even that might cross institutional boundaries around originality and authorship. This confusion reflects a broader issue: many university policies remain vague or outdated when it comes to AI use. As Chiu, et al. [28] and Green, et al. [5] have pointed out, traditional frameworks for academic integrity are ill-equipped to handle the grey areas introduced by generative AI. What emerged clearly was a student desire for ethical guidance, not just rule enforcement. Participants wanted their universities to move away from purely punitive models and adopt more proactive, educative approaches. They argued for clear, accessible resources on how to use AI responsibly, workshops to explore ethical grey zones, and classroom discussions that engage students in reflecting on the broader implications of AI-generated content. In this regard, institutions need to evolve, ensuring that policies not only uphold academic standards but also foster ethical literacy and decision-making. Concerns about data privacy also surfaced, with students unsure about what information ChatGPT might collect and how it is used. These anxieties are not unfounded, given the increasing scrutiny around AI ethics and data governance [21]. Universities must not overlook this dimension; developing robust, transparent data use protocols is just as important as academic policy reform. 5.1.3. Bridging the AI Literacy Gap: Building Competence and Confidence Another recurring theme was the lack of formal training or instruction around AI tools. Many students described learning to use ChatGPT through trial and error, peer recommendations, or social media tutorials. While this organic learning has its merits, it also leaves significant gaps in 455 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate understanding. Without structured guidance, students may misuse AI, either by accident or misunderstanding, and remain unaware of its limitations. This study supports the growing call for comprehensive AI literacy programs in higher education [37]. Such programs should move beyond surface-level tool training and incorporate critical digital literacy, algorithmic awareness, and ethical reasoning. Universities can consider embedding AI literacy within first-year seminars or interdisciplinary courses, ensuring students gain both the technical and reflective skills needed to engage responsibly. Given the accelerating role of AI across sectors, these competencies will be essential not just for academic success, but for future employability and citizenship [38]. 5.1.4. Protecting Human-Centered Learning: AI as a Complement, not a Replacement Several students voiced deeper concerns about how AI might alter the nature of learning itself. Specifically, they worried that heavy reliance on ChatGPT could reduce opportunities for collaboration, discussion, and peer feedback. These concerns reflect long-standing educational theories, such as Vygotsky’s emphasis on the social construction of knowledge [39]. Learning, in this view, is not just an individual act of information processing but a dynamic, dialogic process shaped by social interaction. While AI can offer speed and convenience, it cannot replicate the nuances of human dialogue, the emotional resonance of face-to-face learning, or the collaborative energy of group work. Students noted that some of their most meaningful educational experiences involved debate, exchange of ideas, and shared problem-solving—activities that AI, at present, cannot meaningfully simulate. To address this, educators may consider adopting blended pedagogical approaches that integrate AI in ways that enhance, rather than replace, human learning. For instance, AI could be used to generate prompts for classroom discussion, help students prepare for debates, or offer feedback before peer review sessions. This model retains the value of AI while maintaining the irreplaceable benefits of human interaction [19, 36]. 5.1.5. Broader Reflections and Implications In sum, the discussion points to a shifting educational landscape, where students are actively negotiating the boundaries of what it means to learn in an AI-rich environment. Their perspectives suggest that the integration of tools like ChatGPT is not simply a matter of adoption, but adaptation— requiring thoughtful changes to pedagogy, policy, and institutional support systems. While students clearly see the potential of AI to democratize access, reduce academic stress, and personalize learning, they are also wary of its ethical implications and unintended consequences. Their insights suggest that successful AI integration depends on more than technological access—it requires a culture of ethical awareness, institutional clarity, and ongoing support. Universities stand at a critical juncture. By involving students in conversations around AI use, embedding AI literacy into curricula, and designing policies that prioritize both innovation and integrity, institutions can ensure that generative AI serves as a bridge to deeper learning rather than a barrier to it. 5.2. Conclusion This study offers a deeper, more personal look at how university students perceive and experience using ChatGPT in their academic lives. Through open conversations, it became clear that students see ChatGPT as both a helpful tool and a source of some real challenges—especially when it comes to ethics and the way learning is evolving. Students consistently highlighted the benefits of using ChatGPT, noting how it helps them complete assignments, understand complex ideas, and learn more independently. These positives suggest that AI has real potential to improve learning by making it more personalized, reducing stress, and offering support—especially when instructors can’t always give individual attention. These findings echo what earlier research has pointed out about the role of AI in supporting modern education [20, 36]. 456 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate At the same time, the study also revealed a sense of caution. Some students were worried about relying too much on ChatGPT and losing the ability to think critically or engage deeply with their work. Others shared experiences where the AI gave incorrect or misleading answers. These concerns align with what researchers like Perez and Johnson describe as the "automation paradox"—where convenience might come at the cost of deep thinking [7]. For educators, this raises an important question: how can we use tools like ChatGPT without weakening the intellectual skills we’re trying to build? The answer may lie in thoughtful, balanced integration into teaching [23, 28]. Ethical concerns were a major theme throughout the study. Many students felt unsure about what counts as fair or honest use of ChatGPT. Is it cheating to use it to draft an essay? What if it rewrites your ideas? The lack of clear guidelines made students anxious, especially when it came to issues like plagiarism and privacy. This points to a growing need for universities to create clear, flexible policies that keep up with how fast AI is changing [17, 29]. Students also called for a shift from punishments to education—teaching them how to use AI responsibly instead of just warning them not to. A related issue is that many students admitted they don’t fully understand how ChatGPT works or how to use it effectively. This gap in AI literacy is a barrier to meaningful and ethical use. Without proper understanding, students are more likely to misuse the tool or misunderstand its limits. Embedding AI literacy into the curriculum would give students the skills they need to use these technologies wisely [28, 38]. Another concern was that too much reliance on AI might hurt the social side of learning. Some students worried that if everyone turns to ChatGPT, things like group discussions and peer learning might fade away. But those social interactions are often where deeper learning happens. That’s why teaching methods need to evolve in a way that brings AI into the mix without losing the human connections that make learning meaningful [23, 34]. 6. Recommendations Based on what students shared, the following steps are recommended for universities, educators, and policymakers: 6.1. Develop clear and flexible AI policies Institutions should create transparent guidelines that clearly explain what’s allowed and what’s not when using AI tools like ChatGPT. These policies should evolve over time as AI technology changes, and they should help students make informed choices without fear or confusion. 6.2. Integrate AI Literacy into the Curriculum Students need to understand how AI works, what its strengths and limitations are, and what ethical use looks like. AI literacy programs should be hands-on and practical—not just theoretical—so students can apply what they learn in real-world contexts. 6.3. Promote Ethical Awareness through Education, Not Punishment Instead of focusing on penalties, universities should teach students how to think critically about AI use. Workshops, classroom discussions, and reflective assignments can encourage students to use these tools responsibly and with intention. 6.4. Rethink Assessment Methods To reduce the risk of AI misuse, assessments should focus more on creativity, collaboration, and critical thinking—skills that AI can’t replicate easily. This can help ensure that learning remains authentic and student-driven. 6.5. Adopt Blended Teaching Models 457 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate AI should support, not replace, human-centered learning. Teaching strategies that mix technology with group work, discussions, and experiential learning can keep education both efficient and socially engaging. 6.6. Strengthen Data Privacy and Transparency Universities and AI companies must be clear about how student data is used. Privacy policies should be easy to understand, and students should give informed consent before using AI tools in academic settings. 6.7. Final Thoughts Bringing AI tools like ChatGPT into higher education opens up exciting possibilities—but it also raises important questions. This study gives voice to students who are navigating that space, often without much guidance. Their experiences highlight the need for thoughtful policies, meaningful education, and ethical reflection. By combining clear rules, strong support systems, and inclusive teaching methods, universities can create a space where AI enhances learning instead of replacing it. The goal should not be to resist AI, but to make sure it’s used in ways that deepen understanding, protect academic integrity, and support human growth Transparency: The author confirms that the manuscript is an honest, accurate, and transparent account of the study; that no vital features of the study have been omitted; and that any discrepancies from the study as planned have been explained. This study followed all ethical practices during writing. Acknowledgment: The author would like to express sincere gratitude to her supervisor for her guidance and to all the participants who contributed to this study. Their insights and cooperation were invaluable to the completion of this research Copyright: © 2025 by the author. This open-access article is distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). References [1] M. Abdaljaleel, S. Khan, and A. Farooq, "Student perceptions of ChatGPT as an academic support tool in higher education," Journal of Educational Technology & Society, vol. 27, no. 1, pp. 45–60, 2024. [2] C. Lin and J. Wang, "ChatGPT as a study companion: Autonomous learning in large classrooms," Educational Technology Research and Development, vol. 72, no. 1, pp. 23–39, 2024. [3] A. Johnson, R. Smith, and T. Nguyen, "Adapting instructional design in the age of AI: Beyond automation," Journal of Educational Innovation, vol. 12, no. 3, pp. 212–228, 2023. [4] R. Brown, D. Evans, and L. Nguyen, "Equity and accessibility in AI‑assisted learning environments: Policy considerations," Higher Education Policy Review, vol. 31, no. 4, pp. 333–349, 2024. [5] S. Green, R. Patel, and J. Williams, "AI-generated text and academic integrity: Developing novel detection algorithms," Journal of Educational Computing Research, vol. 61, no. 5, pp. 1193–1210, 2023. [6] A. Bodani, J. Martin, and P. Torres, "The automation paradox: How AI assistance impacts critical thinking skills in university students," Computers & Education, vol. 190, p. 104627, 2023. [7] D. Perez and T. Johnson, "The automation paradox and educational technology: Balancing AI support and cognitive engagement," Journal of Educational Psychology, vol. 115, no. 5, pp. 945–958, 2023. [8] P. Hernández and G. López, "Institutional policy gaps in AI integration for education: A global perspective," Educational Policy Analysis Archives, vol. 32, no. 16, pp. 1–24, 2024. [9] Y. Hou, J. Chen, and M. Lee, "Effects of ChatGPT on peer‑to‑peer engagement in student learning communities," Learning, Media and Technology, vol. 50, no. 1, pp. 54–68, 2025. [10] R. Singh and P. Roy, "Collaborative development of AI policies in education: A multi-stakeholder approach," Educational Policy Analysis Archives, vol. 33, no. 5, pp. 1–29, 2025. https://creativecommons.org/licenses/by/4.0/ 458 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 7: 448-458, 2025 DOI: 10.55214/25768484.v9i7.8592 © 2025 by the author; licensee Learning Gate [11] S. Martin and D. Clark, "Developing living policies for AI in education: Flexibility and innovation balance," Policy Futures in Education, vol. 22, no. 4, pp. 440–459, 2024. [12] A. Jobin, M. Ienca, and E. Vayena, "The global landscape of AI ethics guidelines," Nature Machine Intelligence, vol. 1, no. 9, pp. 389-399, 2019. [13] A. Roberts, P. Singh, and S. Kumar, "Transparency and user control in AI platforms: Implications for student privacy," Ethics and Information Technology, vol. 26, no. 2, pp. 151–167, 2024. [14] X. Zhang, Y. Liu, and M. Chen, "ChatGPT for critical reading and research: Benefits and risks," Journal of Educational Computing Research, vol. 62, no. 6, pp. 1431–1450, 2024. [15] M. Bali, ChatGPT in education: Hype, harm, or help? USA: Hybrid Pedagogy, 2023. [16] M. O’Neill and S. Patel, "From punishment to education: Shifting academic integrity approaches in the AI era," Higher Education Ethics Quarterly, vol. 19, no. 1, pp. 30–46, 2024. [17] Y. Wang and L. Thomas, "Redesigning assessments for the AI age: Emphasizing reflection and personalization," Assessment & Evaluation in Higher Education, vol. 49, no. 3, pp. 409–425, 2024. [18] K. Evans and J. Matthews, "AI literacy in higher education: Preparing students for ethical AI use," International Journal of Educational Research, vol. 117, p. 102098, 2023. [19] A. Kumar and P. Singh, "Data privacy challenges in AI‑assisted learning: Student perspectives," Journal of Information Ethics, vol. 33, no. 1, pp. 14–27, 2024. [20] M. Lee, Y. Zhang, and K. Hernandez, "Leveraging AI for inclusive education: ChatGPT and students with disabilities," Disability Education Journal, vol. 29, no. 1, pp. 78–92, 2025. [21] MIT AI Policy Review, "Preliminary guidelines for ethical AI use in education," MIT AI Policy Review, vol. 2, no. 1, pp. 10–28, 2024. [22] R. Nguyen, R. Brown, and Q. Zhao, "Building AI literacy for ethical AI use in higher education," Journal of Educational Computing Research, vol. 62, no. 3, pp. 867–889, 2024. [23] L. Nguyen, B. Johnson, and K. Evans, "Fostering peer collaboration alongside AI integration in higher education," Journal of Computer Assisted Learning, vol. 41, no. 1, pp. 103–120, 2025. [24] A. Shafqat and S. Amjad, "Perceptions of ChatGPT in Pakistani universities: Benefits and policy challenges," Pakistan Journal of Educational Research, vol. 39, no. 1, pp. 77–95, 2024. [25] R. Sharma and S. Amjad, "Perceptions of ChatGPT in Pakistani universities: Benefits and policy challenges," Pakistan Journal of Educational Research, vol. 39, no. 1, pp. 77–95, 2024. [26] Q. Zhao and M. Lee, "Integrating AI literacy into higher education curricula: A framework and case studies," Computers & Education, vol. 188, p. 104613, 2024. [27] D. Baidoo‑Anu and L. Owusu Ansah, "Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and learning," Journal of AI, vol. 7, no. 1, pp. 52–62, 2023. https://doi.org/10.61969/jai.1337500 [28] T. K. F. Chiu, J. C. H. Lee, and C. K. Y. Mak, "Students’ perceptions and usage of ChatGPT in higher education: A survey study," Education and Information Technologies, pp. 1–15, 2023. [29] T. Baker and L. Smith, "Challenges of detecting AI-generated academic content: A review of plagiarism tools and policy implications," Journal of Academic Integrity, vol. 18, no. 2, pp. 101–115, 2024. [30] F. D. Davis, "Perceived usefulness, perceived ease of use, and user acceptance of information technology," MIS Quarterly, vol. 13, no. 3, pp. 319-340, 1989. https://doi.org/10.2307/249008 [31] M. Hernandez and S. Kim, "Meta-analysis of AI tools in higher education: Efficiency, engagement, and pedagogy," Review of Educational Research, vol. 95, no. 1, pp. 85–110, 2025. [32] B. Johnson, S. Wang, and M. Clark, "Social dynamics and AI in education: The impact on student motivation and belonging," Educational Psychology Review, vol. 36, no. 2, pp. 567–586, 2024. [33] S. Kumar, V. Patel, and H. Lee, "AI integration in online classrooms: Balancing technology and human interaction," Computers in Human Behavior, vol. 136, p. 107385, 2024. [34] D. Lee and Y. Chen, "Disciplinary differences in AI use: Student attitudes toward ChatGPT," Journal of Learning Analytics, vol. 11, no. 2, pp. 49–64, 2024. [35] R. Luckin, W. Holmes, M. Griffiths, and L. B. Forcier, Artificial intelligence and education: Promises and implications for teaching and learning. London, U.K: Routledge, 2024. [36] D. Cotton, P. Cotton, and J. Shipway, "Chatting and cheating? Exploring the role of ChatGPT in student academic integrity," Innovations in Education and Teaching International, vol. 61, no. 2, pp. 228-239 2023. https://doi.org/10.1080/14703297.2023.2190148 [37] R. Luckin, W. Holmes, M. Griffiths, and L. B. Forcier, Intelligence unleashed: An argument for AI in education. London, U.K: Pearson Education, 2016. [38] P. Smutny and P. Schreiberova, "Chatbots for learning: A review of educational chatbots for the Facebook Messenger," Computers & Education, vol. 151, p. 103862, 2020. https://doi.org/10.1016/j.compedu.2020.103862 [39] E. Kasneci et al., "ChatGPT for good? On opportunities and challenges of large language models for education," Learning and individual differences, vol. 103, p. 102274, 2023. https://doi.org/10.1016/j.lindif.2023.102274 https://doi.org/10.61969/jai.1337500 https://doi.org/10.2307/249008 https://doi.org/10.1080/14703297.2023.2190148 https://doi.org/10.1016/j.compedu.2020.103862 https://doi.org/10.1016/j.lindif.2023.102274