Document Type : Original Article
Authors
1
Assistant Professor, Department of Information Sciences & Knowledge Studies, University of Tabriz, Tabriz, Iran
2
Ph.D Student in Knowledge & Information Science, University of Tabriz, Tabriz, Iran
3
Lecturer, Sociologist and Social Researcher, Member of the National Foundation of Iranian Elites, Tehran, Iran
Abstract
Introduction: In recent years, artificial intelligence (AI) has emerged as a transformative force in higher education, reshaping established approaches to learning, research, and knowledge production. Graduate students, as active contributors to scholarly knowledge production, are increasingly incorporating AI-based tools into their educational and research activities. Applications such as ChatGPT, Copilot, Gemini, DeepSeek, and other intelligent assistants have expanded opportunities for academic writing, information retrieval, data analysis, idea generation, programming, and learning. However, the rapid advancement and widespread adoption of generative AI have also raised concerns regarding scientific accuracy, the generation of fabricated or unreliable references, research ethics, academic authenticity, and the credibility and reliability of AI-generated content. Given the growing integration of these technologies into academic practices, understanding how graduate students use AI tools, the benefits they perceive, and the challenges they encounter is essential for informed educational planning and evidence-based policy development. Accordingly, this study aimed to investigate patterns of AI tool use in the educational and research activities of graduate students at the University of Tabriz and to identify the opportunities and challenges associated with such use.
Methodology: This qualitative phenomenological study involved 18 graduate students from diverse academic disciplines who were recruited through purposive snowball sampling until theoretical saturation was achieved. Data were collected using semi-structured interviews and analyzed through open, axial, and selective coding based on the approach developed by Glaser and Strauss. Data collection and analysis were conducted concurrently to facilitate the refinement of emerging categories and enhance the trustworthiness of the findings.
Findings: The findings indicated that AI tools have become increasingly integrated into the academic activities of graduate students, although the frequency and nature of their use varied according to disciplinary background and the stage of the academic or research process. Participants reported using AI across a broad range of tasks, including academic writing, research proposal development, literature reviews, preparation of educational content and learning materials, data analysis, manuscript preparation, programming, idea generation, and reference-related activities. Research-related activities constituted the most prominent area of AI use, particularly for academic writing and data analysis. In educational contexts, students reported using AI to facilitate learning, clarify complex concepts, prepare and organize instructional materials, and support self-directed learning.
Participants identified several significant benefits associated with the use of AI tools. These tools were perceived as reducing the time required to complete academic tasks, facilitating academic writing and data analysis, helping students overcome certain skill-related limitations, enhancing motivation for research, and supporting independent learning. Some participants also viewed AI as a valuable resource for generating ideas and fostering critical thinking, particularly when its outputs were evaluated critically rather than accepted uncritically. Overall, these findings suggest that AI can serve as a complementary academic resource that enhances students’ capabilities and extends their capacity to engage with academic tasks, rather than merely automating discrete activities.
However, participants also identified several substantial challenges associated with the use of AI tools. Concerns regarding the accuracy, reliability, and scientific validity of AI-generated information were particularly prominent. Students reported encountering scientific errors, inaccurate or misleading information, fabricated references, and unreliable sources. Consequently, most participants emphasized that AI should be regarded as a “research assistant” rather than as a substitute for the researcher. They described verifying AI-generated information against authoritative and multiple sources, comparing AI outputs with their prior knowledge and experience, critically evaluating the results, and conducting independent assessments before incorporating them into their academic work.
The findings also revealed significant access and infrastructure-related barriers. Internet filtering and other restrictions affecting access to certain AI services, sanctions, the high cost of advanced subscriptions, and limitations of free versions were identified as factors that could constrain effective and sustained use. Disciplinary differences were also evident. Students in humanities-oriented fields tended to use AI more extensively for academic writing and language-related tasks, whereas students in technical and quantitative disciplines reported greater use for data analysis, programming, and coding. Overall, the findings indicate a pattern of conditional and critical adoption, whereby students recognize the practical value of AI while remaining aware of its limitations and the continued need for human judgment, critical evaluation, and verification.
Discussion and Conclusion: The study concludes that AI is no longer merely an emerging technology but is increasingly becoming integrated into the everyday educational and research practices of graduate students. Its contributions are particularly evident in enhancing the speed and efficiency of academic tasks, reducing certain skill-related barriers, facilitating learning and academic writing, strengthening research motivation, and supporting analytical and technical activities. Nevertheless, the effective use of AI depends largely on students’ ability to critically evaluate AI-generated content and integrate useful outputs with credible and authoritative scholarly sources. Two major categories of challenges—access and infrastructure constraints, and concerns regarding accuracy and authenticity—may undermine the quality and reliability of AI-supported academic activities. Accordingly, universities should facilitate equitable access to appropriate AI tools, promote the development of academic AI literacy, and establish mechanisms for verifying AI-generated information, sources, and references. Educational and training programs should place particular emphasis on critical evaluation of AI outputs, detection of misinformation and fabricated content, responsible and ethical use of AI, and transparent disclosure of AI assistance in scholarly work. Overall, the findings provide a basis for universities to develop evidence-based policies and educational strategies aimed at the responsible, effective, and sustainable integration of AI into graduate education and research.
Keywords
Subjects
