Generative artificial intelligence and reference management tools have become embedded in the everyday assignment writing of students at Zimbabwe Open University, according to a qualitative case study published in Frontiers in Artificial Intelligence. The research, conducted at the university’s Masvingo Regional Campus within an Open and Distance eLearning (ODeL) environment, found that tools such as ChatGPT, Grammarly, Mendeley, Zotero and ProWritingAid enhance writing quality, research efficiency, accessibility and learner confidence, while simultaneously raising concerns about overreliance, academic integrity and diminished critical engagement.
The study, led by Logic Magwa of the Institute for Open Distance Learning at the University of South Africa, adopted a qualitative instrumental case study design. Participants were selected through purposive sampling based on their experience with AI-supported learning. Data were collected through semi-structured interviews lasting between 45 and 60 minutes with seven lecturers and one quality assurance coordinator, and through focus group discussions of 70 to 90 minutes with fourteen students, with two student representatives drawn from each faculty. The research was guided by Constructivist Learning Theory, which holds that knowledge is actively constructed by learners through interaction with their environment rather than passively received.
Analysis followed Braun and Clarke’s six phases of thematic analysis, with transcripts read and re-read, coded, and refined into three overarching themes: academic support and writing enhancement; research and content development; and practical efficiency and accessibility. Ethics approval was obtained from the University of South Africa’s College of Education Research Ethics Committee, and permission for fieldwork was granted by Zimbabwe’s Ministry of Higher and Tertiary Education. Member checking was conducted after analysis, and trustworthiness criteria of dependability, confirmability, transferability and credibility were applied.
The first theme captured how generative AI improves grammar, clarity, structure and learner confidence. One lecturer described AI tools as “a rich knowledge reserve” through which students access desired information easily and without wasted time. Another noted that AI tools serve as models of standard academic English, though cautioned that students who rely heavily on AI-generated language may reproduce polished text without fully understanding the underlying concepts, weakening critical engagement and authentic voice development. A student recounted benefiting greatly from AI tools during teaching practice at a school with no library, completing assignments effectively with AI resources.
The quality assurance coordinator offered a balanced institutional perspective, observing that AI-generated content provides a useful starting point but that beginning with AI-produced drafts may reduce the cognitive struggle associated with constructing ideas independently, potentially shifting AI from a scaffold within the Zone of Proximal Development to a substitute for active knowledge construction. This tension between efficiency and authentic learning ran through all participant groups: students emphasised immediate support and accessibility, lecturers acknowledged benefits while worrying about overreliance, and the coordinator framed AI as helpful only when it does not replace independent cognitive effort.
The second theme concerned research and content development. Participants reported that AI tools assist in brainstorming, summarising articles, gathering educational theories and locating research articles relevant to a student’s area of study. One lecturer, however, expressed concern that students may rely on AI-generated summaries rather than engaging directly with primary scholarly sources, limiting their exposure to nuanced arguments and methodological diversity. The quality assurance coordinator similarly warned that excessive dependence on AI-assisted summarisation may compress complex academic debates into simplified outputs, reducing opportunities for critical analysis and independent synthesis.
The third theme highlighted practical efficiency and accessibility. Students described accessing information easily on mobile phones wherever there is network coverage, and noted that AI provides access to comprehensive educational resources that may otherwise be difficult to obtain. Lecturers reported that automatic editing and formatting allow student teachers to concentrate on content and save time, with tools such as Zotero used for formatting, ChatGPT for drafting and ProWritingAid for editing. Yet one lecturer questioned whether such efficiency obscures the depth of students’ cognitive engagement, making it difficult to determine whether improvements reflect genuine skill development or technological assistance. Some lecturers also reported difficulty distinguishing authentic student work from AI-assisted submissions.
Interpreted through the constructivist framework, the findings suggest that generative AI functions as a cognitive scaffold supporting learners within their Zone of Proximal Development, but that when learners depend excessively on AI-generated outputs without critically interrogating or adapting them, scaffolding shifts toward cognitive substitution. The discussion notes that in ODeL contexts, where lecturer-student interaction is limited, AI tools may inadvertently replace the dialogic feedback processes central to constructivist pedagogy, making feedback automated and transactional rather than negotiated. The author also cautions that generative AI systems do not access live databases in real time and may produce incomplete, biased or fabricated references, so overreliance on AI-generated literature may narrow scholarly engagement in environments where access to subscription databases is already limited.
The study situates itself within a research gap: most prior studies of AI in academic writing were conducted outside Zimbabwe, often in well-resourced conventional universities, and existing Zimbabwean studies addressed general academic writing rather than assignment-specific processes in teacher education or ODeL contexts. Teacher education was singled out as a critical site because student teachers are expected to transfer learned academic skills into classroom instruction, meaning their AI use has implications for how future educators may integrate or regulate AI in their own teaching. The findings align with international research reporting AI-driven gains in grammatical accuracy, coherence and workflow efficiency, and with Zimbabwean studies reporting enhanced research efficiency alongside accountability concerns.
The author recommends that institutions develop comprehensive AI policies defining acceptable and unacceptable uses of generative AI, introduce structured AI literacy programmes for students and lecturers, and adjust curriculum design to incorporate reflective and process-based assessment methods such as oral defences, reflective commentaries and staged submissions. The study’s limitations are acknowledged: findings rest on self-reported data subject to recall bias, were gathered at a single institutional site limiting generalisability, and analysis by a single coder may introduce subjectivity, although reflective memo-writing was used to enhance consistency. Future research, the paper concludes, should consider multi-site studies, multiple coders and mixed methods approaches.
Subject of Research: A qualitative case study of how generative AI and reference management tools are integrated into assignment writing at an Open and Distance eLearning university in Zimbabwe.
Article Title: Integrating generative AI and reference management tools in assignment writing: insights from an ODeL context
Article References: Magwa, L. (2026). Integrating generative AI and reference management tools in assignment writing: insights from an ODeL context. Frontiers in Artificial Intelligence, 9, Article 1954585. https://doi.org/10.3389/frai.2026.1954585
Image Credits: AI Generated
DOI: 10.3389/frai.2026.1954585
Keywords: generative AI, ChatGPT, Grammarly, Mendeley, academic writing, Open and Distance eLearning, Zimbabwe Open University, assignment writing, academic integrity, constructivist learning theory, teacher education, qualitative case study
News Source: Denise Maddox. (October 29, 2026). Generative AI Reshapes Assignment Writing at Zimbabwe’s Open University, Study Finds. Scienmag.
