Developing student writing skills in universities:

Is GenAI an aid or a hindrance?

John Ebejer University of Malta john.ebejer@um.edu.mt Introduction – Why this essay? Much is being said and written about the benefits that AI brings to the workplace. Similarly, many express concerns about the negative impacts that reliance and inappropriate use of AI has on skill development particularly in universities. This essay focuses on the impact of GenAI on the development of students’ written communication skills. I contend that student use of GenAI needs careful management by academics and university authorities to bring about the maximum benefits to the learning process. Various online news portals claim that AI is being used extensively by students in many universities (Bloomberg, 2025). A survey on academic integrity in higher education in the UK found that thousands of students were caught misusing ChatGPT and other AI tools in recent years (Goodier, 2025). In 2023-24, five for every thousand students were found to have cheated using AI tools. Experts claim that these findings are the tip of an iceberg. An educator at a high school on British Columbia claimed that there has been an incremental increase in student usage of AI until this year when it just “exploded” (Perkel, 2025). He claims that “literally all students are using it this year.” I have been teaching at tertiary level full time for the past twelve years. This past year, I assessed numerous assignments as well as several undergraduate dissertations. It is evident that several students used AI. Some used it in a constructive way like (for example improving on the writing quality of the student’s text). Others use it abusively to generate

text for their assignment bypassing any effort to critically think about the assignment question. I am coordinator for the undergraduate tourism management programmes at my university. As such, I am concerned about the inappropriate use of GenAI because it undermines student learning and hampers the development of writing skills. Prior to becoming an academic, I worked professionally for twenty years. Therefore I bring to the debate a unique perspective with a mix of professional and academic work experience. I worked as an architect, an urban planner, a project coordinator and a tourism consultant. I also held senior management positions in public sector organisations. My work experience gives me useful insights on the importance of writing skills in a work environment. In addition, my teaching experience gives me some idea how students perceive GenAI and the inevitable pressures for students to use it. The purpose of this essay is twofold. In the first instance I would like to raise awareness about the need for university students to develop their writing and critical thinking skills. Second I discuss how GenAI effects learning processes in universities particularly in relation to the essential skill of written communication. For a meaningful debate, we need to draw a distinction between different forms of AI. Traditional AI refers to systems designed to respond to a particular set of inputs. It performs specific tasks intelligently (Marr, 2023). There are a broad range of AI tools that facilitate complex tasks such as computation, drawing and processing of data. Some professions such as engineering and architecture have been using AI tools for decades. In tertiary education, the use of traditional AI tools is overwhelmingly positive and should be encouraged. On the other hand GenAI is a form of AI that creates something new. GenAI systems, such as ChatGPT and Gemini, are powered by Large Language Models (LLMs). They generate human-like text by scraping data from the internet. There are also Assistive AI tools. These suggest, correct and improve text written by the student (Sage, n.d.). Making these distinctions is useful because there is some literature that attributes the benefits of traditional AI to GenAI thus confusing the issues.

The need for skilled writers at the workplace The primary objective of higher educational institutions is to prepare students to enter the workforce (Southwork et al, 2023). Universities seek to develop the skills and knowledge that students will eventually require when they start working as professionals. In the first instance managers and professionals need to be knowledgeable in their area of expertise. In terms of soft skills, top of the list are writing skills and critical thinking. When recruiting, written communication and problemsolving skills feature strongly in employers’ list of priorities (Nilson, 2023). Experienced managers often lament the writing skills of persons seeking entry-level positions (Wise, 2005). This is also my experience. In my professional work experience, I came across many people at different management levels who were poor at written communication. Writing and thinking skills distinguish the competent from the mediocre. The absence of writing skills is detrimental to organisational efficiency. Hargie et al (2004) list various forms of written communication that take place in business organisations including “audit reports, shareholder statements, marketing and promotional materials, annual reports, technical briefs, white papers and other forms of writing that interface with services and products.” They note that most business managers spend between eight and fourteen hours per week in writing-related activities such as “producing disciplinary action reports, clarifying job procedures, dealing with formal grievances, writing memos, producing instructional documents to subordinates, drafting incident reports and writing external letters or reports to customers.” In addition, they spend significant amounts of time responding to written communications of others. Similarly in my professional working years, I wrote hundreds of reports, policy documents, draft legislation, development briefs, technical reports, annual reports and progress reports. I also wrote thousands of work-related emails. Looking back would I have used LLMs to generate text for any of my reports or emails, if the tool were available? Certainly not. Every report referred to a context and a set of circumstances including site features, people’s expectations, technical issues and implications on expenditure. It could all get very complex. When writing, the overall context was foremost in my mind and therefore my writing could not have been replicated by AI.

Professional work requires the skill and knowledge of a competent professional. Managers and professionals are held responsible for the decisions that they take. Managers are required to make decisions taking into account the overall context, including technical issues. They are required to evaluate and make judgements based on reliable information. A broad range of social, environmental, cultural, management and/or political issues may need to be considered – issues that should be handled by professionals experienced in critical thinking and communication. In all jobs there is the people factor; these being employees, clients or external stakeholders. The context within which decisions are taken can get exceedingly complex. It would be downright irresponsible for a manager to take decisions on the basis of data scraped from LLMs. Some commentators claim that AI will radically change work processes of managers and other professionals (such as accountants, lawyers, architects, doctors, etc.). Because of the media hype on AI, many young people are led to believe that they will get by in their careers using generative and other forms of AI. This is misleading. Even worse, it makes them complacent. They are less likely to strive to enhance their writing and critical thinking. Another important consideration is that a writer improves the quality of their writing with every written communication. Today I consider myself a good writer. That came about because I dedicated time and effort in everything I wrote. Each writing task was an opportunity to improve my skill. Whether at the workplace or in academia, effective writing involves an ongoing process of drafting, redrafting, editing and refining. Let’s take this essay as an example. I could have used GenAI to generate a first draft but this would have thrown me in a direction that I might not have wished for. I chose NOT to use AI to ensure that the final outcome would reflect precisely the thoughts and ideas that I have developed over recent months and years about the subject. I drafted new text. I redrafted sentences and paragraphs multiple times. I reordered the paragraphs into a new structure. It took me loads of time. But finally, it is exactly the outcome that I intended. In my professional work, I collaborated with many people, mostly managers or professionals like myself. I noted that the best achievers were those who were effective in their written communication. With

their reports and emails they were able to communicate their thoughts and ideas in a convincing manner. Of course, verbal skills in presentations and meetings are also important in the workplace but it is skilful written communication that makes managers most effective. I also observed that good writers are good at critical thinking. They are able to break down an issue or problem into its components parts and come up with effective solutions by careful analysis of each component. Writing is closely associated with critical thinking. In writing a work report, a manager is reorganising information and ideas. They present them in an understandable manner in written format. My professional work experience made me acutely aware of the need for new graduates to have good writing skills. For this reason, one of my objectives in university teaching is to enable students to enhance their writing and critical thinking skills. This is to prepare them for their eventual positions in middle and senior management. I am passionate about the importance of writing skills. I am so not because I read it somewhere but because I lived it for twenty years in my professional career. Now I continue to live it in academia. Writing – not just stringing sentences together Writing is not just about putting a string of sentences together. It is about the ability to communicate thoughts, ideas, concepts and information to a target audience. The writing style of a progress report for a management board is different from a technical report for professional colleagues. In written communications, precision is essential. Every detail matters – the report structure, the sequencing of paragraphs, the structure of sentences and the choice of words and phrases. In an academic context, the writer must refer to reliable sources that are appropriate for the task at hand. The internet throws up a huge number of potential sources and it is up to the student to use filters and their own judgement to find the ones that are most useful. The student has to read and understand the relevant sections of the selected sources. Then comes the writing. This is a time-consuming process but one that enables the student to learn about the subject and more importantly to develop their skills of writing and critical thinking.

With GenAI, a student can short-circuit the process and produce an assignment that apparently responds to the assignment question and also includes references. The student will get the grade but will miss out on the learning. In particular they will miss out on developing their writing skills. At my university, I teach Critical Thinking and Communication and I have been doing so for several years. I have always insisted with students that if they want to improve their writing skills they need to read. Normally the writing of a good essay or report requires a lot of reading. It is through reading that the student develops a sound understanding of the subject. Calls to “embrace” GenAI are an implicit encouragement to students to read less. By using GenAI to write their essays, students end up doing little or no reading and this is detrimental to their learning. They are also missing out on the opportunity to practice their writing. The purpose of assignments is to develop student skills including writing, reading, critical thinking and carrying out literature searches. For learning to be achieved there needs to be understanding. The use of ChatGPT provides a short cut for the student to produce the output but in so doing the student is reducing the understanding and hence the learning. The pros and cons of student use of GenAI In an ideal world all students would use GenAI ethically and in a manner that enhances their learning. In the real world, many students are subject to time and other constraints. Many students take-on parttime jobs to support themselves financially during their studies. Using GenAI to write their assignments is too much of a temptation for many. This enables them to carry out their assignments at a fraction of the time. Teachers are also subject to constraints. Some may be unwilling or unable to dedicate the extra time that is required to assist students to use GenAI in an ethical manner. Inappropriate use of GenAI by students results in reduced learning of skills that are essential for their eventual professional work, particularly critical thinking and written communication. When students are requested to write essays or report, they are using their thinking and writing skills – skills that can only be developed through practice, practice and more practice. If students repeatedly use AI to generate text,

they deprive themselves the opportunity to develop these essential skills. Comparing with the world of sports, any sports person knows that to achieve success they need to practice to improve the various body skills that are required for their sports. It is the same with writing. Practice is needed to enhance the mentals skills that are required to become a good writer and communicator. Gen AI has created a new reality in which it is much more difficult for students to develop their writing skills. It is also difficult for teachers to create assignments and class exercises that help students in their writing. A study by researchers at MIT suggests that use of LLMs could actually harm learning, especially among younger users (Chow, 2025).^1 The research suggests that over the course of several months, regular users of ChatGPT got lazier with each subsequent essay, dedicating less time to the thinking that is normally required for writing an essay. Interestingly I came across a debate on social media about this study. From comments made, GenAI campaigners insist that students should be encouraged to use GenAI because they will eventually need it at the workplace. However they are unwilling to acknowledge that widespread use of GenAI undermines learning processes in universities. Nor are they willing to concede that, for some disciplines, it is difficult to design assignments that safeguard against the abusive use of GenAI. Some studies claim that ChatGPT enhances skill development (Daniel, et al, 2025; Boubker, 2024). They are largely based on the comparison of academic work using ChatGPT with other work which is done without the use of AI. The output is noticeably better when ChatGPT is used. But it is incorrect to assume that an improved output reflects improved learning. Deng et al (2024) note that text revised by AI does not reflect the student’s writing competence. It gives the misleading impression that the student is a proficient writer. Whilst acknowledging its potential for content generation, many educators have expressed concern about academic integrity, overreliance on technology and the potential negative impacts on

(^1) The research by MIT’s Media Lab is yet to be peer reviewed. The researchers decided to go public before publication to inform policy makers at a time when important decisions on education and AI are being taken.

essential skills such as writing and problem-solving (Deng et al, Wise et al, 2024). Ausmeier (2025) highlights some of the fears of educators and students at the University of Maastricht. Educators are concerned that learning is compromised by GenAI usage. On assessments, these tools enable students to take shortcuts to avoid carrying out the work that is expected of them. University of Maastricht academics also worry about the decreased quality and potential fraud. On the other hand, students feared being incorrectly accused of AI fraud. They also expressed concern that that their education might not provide them with adequate skills in a future work environment that is dominated by AI. Recently the University of Malta organised a debate on the subject. The subject was debated by a panel that included academics from Norway, Croatia and Malta (Keszthelyi, 2025). There was general agreement that GenAI is an important development in university teaching and efforts should be made for teachers and students to make effective use of it. There were also difference of opinions. Some academics downplayed the risks of abusive use of GenAI whereas others expressed concern that excessive use of GenAI would be significantly detrimental to student learning. A pertinent question is how can GenAI influence the reliability of assessment methods? For reports, essays and take-home assignments, there is a high risk that GenAI is used by the student without consent (University of Twente, 2024). In the case of project work and presentations, there is also the risk of unauthorised usage depending on the task. This raises the issue of fairness – for assessments where the use of GenAI is not allowed, students who use it gain an unfair advantage over those who stick by the rules. In my teaching, I encourage students to take responsibility for their own learning. I encourage them to do their own writing because it is through practice that they can develop the skill of writing. I am conscious of the fact, however, that they may end up getting a lower grade then their colleagues who use AI to generate text. Even if there are risks, GenAI can also be used constructively for learning and also to improve efficiency. For example, a student who struggles to be creative could use AI to brainstorm and come up with an idea that best suits the task at hand. GenAI could also be useful to students whose native language is not English. They could improve their English by running their own text through GenAI and then compare the

AI-generated text with their own original draft. They learn by noting how GenAI improved the various paragraphs and sentences. Another possible use of GenAI is to identify sources for essays. There is the risk however that GenAI throws up mostly unreliable internet sources. This was evident in some of the essays that I assessed in recent months. It is preferable for students to learn how to carry out searches on the internet. Even if it takes more time, it enables them to choose the most relevant and reliable sources. ChatGPT at the workplace There are some who claim that ChatGPT will revolutionise the workplace. This is because the tool will be used extensively in many workplaces and in all sectors (Borg, 2024). But is this true? According to Chuma & De Oliveira (2023), ChatGPT is useful for business organisations for compiling research, recommendations on investment decisions, drafting marketing content, brainstorming ideas, delivering aftercare services and increasing customer engagement. A lawyer friend of mine explained that his firm is exploring ways to use GenAI. Efficiency could be improved by using it to for the more mundane tasks such as legal research, contract analysis and document review. That releases the lawyers to work on more value-added tasks such as strategic thinking, problem solving and decision making. On the other hand, the manager of a small accountancy firm claims that his office does not use GenAI because it has no place in the specialised work of accountancy. According to one report, there are many writers who are employed to fix text churned-out by AI. For example a product marketing manager was approached by a content agency to rework website text produced by GenAI. Rather than making small changes, she had to rewrite the entire copy. She explained that AI generated text “was supposed to sell and intrigue but instead it was very vanilla” (Bearne, 2025). Depending on the sector, GenAI can be useful tool at the workplace but only if the user is well-skilled in writing and thinking skills. In the academic literature, there is not much research on GenAI in the workplace. Management research could provide tentative replies to

pertinent questions: To what extent is GenAI being used at the workplace? What role does GenAI have on decision taking of managers and professionals? With a constantly evolving technology, what future role is GenAI likely to have in different industrial sectors? Conclusion the need to adapt In spite the guidelines issued by universities, educators and students are at the receiving end of confused messaging. Students are encouraged to learn how to use GenAI to prepare themselves for the workplace. They are also discouraged from using it for assignments. Where they are allowed to use it, they are expected to do so ethically. Educators have no means to ensure that students use GenAI ethically so compliance is down to the goodwill and ethical standards of individual students. Many will comply but some will not. This makes it difficult to have a level playing field in the classroom. The use of GenAI has brought about new realities in universities. To maximise student learning, educators need to adapt their teaching and assessment method. This is essential if universities are to strengthen their role as places for learning. There is a need for assessments to move away from essay-type questions to assignments that are project-based or that involve problem solving. However for some subjects, there is a limit to how many realistic project-based assignments can be developed. There is also the option of increasing the role of exams in the assessment process. The challenge for universities is how to harness the potential of GenAI while addressing its many challenges. Vrågård et al (2024) recommend the development of clear guidelines on the proper use of such digital technologies. Rather than an outright ban on GenAI, they recommend increased awareness of the benefits and risks amongst educators and students. For universities, one possible approach to GenAI is to shift towards exams in a controlled environment and the adoption of formative assessment that are less vulnerable to responses generated by AI. The educator should consider redesigning the assessment method to make it more resilient against the use of LLMs. There may also be the need to rethink the learning objectives (University of Twente, 2024).

Since the introduction of GenAI in 2022, the situation in universities has evolved and it will continue to evolve for the foreseeable future. Whatever the eventual outcomes on AI, universities must remain focused on creating a learning environment in which students develop the skills of written communication and critical thinking. References Ausmeier, R. (2025, February) What do we really value? Rethinking our approach to education at UM in the age of GenAI. EDLAB, Maastricht University. Available at https://www.maastrichtuniversity.nl/news/edlab/eduminded/what-dowe-really-value-rethinking-our-approach-education-um-age-genai (Accessed on 21 June 2025). Bearne, S. (2025, July 4) ‘I’m being paid to fix issues caused by AI’. BBC. Available at https://www.bbc.com/news/articles/cyvm1dyp9v2o Accessed on 25July 2025. Borg, E ( 2024, November 2) Drop taboos around AI and help students learn to use it. The Times of Malta. Available at https://timesofmalta.com/article/droptaboos-around-ai-help-student-learn-use-it.1100457. Accessed on 25 July 2025. Boubker, O. (2024) From Chatting to Self-Educating: Can AI Tools Boost Student Learning Outcomes? Expert Systems With Applications 238: 1820. https://doi.org/10.1016/j.eswa.2023.121820. Bloomberg Editorial Board (2025, May 27) Does College Still Have a Purpose in the Age of ChatGPT? Bloomberg. Available at https://www.bloomberg.com/opinion/articles/202505 27/ai-role-incollege-brings-education-closer-to-a-crisis-point (Accessed on 28 May 2025). Chow, A.R. (2025, Jun 17) ChatGPT May Be Eroding Critical Thinking Skills, According to a New MIT Study. Time. Available at https://time.com/7295195/aichatgpt-google-learningschool/?fbclid=IwY2xjawLF10hleHRuA2FlbQIxMAABHsxxaQKgDLrOzC6DC5LL-mBinFEzg7R1gXhKQ_HDGapeGslEO62XBzAx6F_aem_Pxht2IgQVvWyd4GU__N7YQ. Accessed on 22 June 2025. Chuma, E. L., & De Oliveira, G. G. (2023). Generative AI for business decision-making: A case of ChatGPT. Management Science and Business Decisions , 3 (1), 5-11. Daniel, K., Msambwa, M. M., & Wen, Z. (2025). Can Generative AI Revolutionise Academic Skills Development in Higher Education? A Systematic Literature Review. European Journal of Education , 60 (1), e70036. Deng, R., Jiang, M., Yu, X., Lu, Y., & Liu, S. (2024). Does ChatGPT enhance student learning? A systematic review and meta-analysis of experimental studies. Computers & Education , 105224. Goodier, M. (2025, June 15) Revealed: Thousands of UK university students caught cheating using AI. The Guardian. Available at https://www.theguardian.com/education/2025/jun/15/thousands-of-ukuniversity-students-caught-cheating-using-ai-artificial-intelligence-survey. Accessed on 20 June 2025.

Gudge, E. (2025, June 05) ‘Has AI ‘transformed’ university for the better?’ BBC News, Oxford. Available at https://www.bbc.com/news/articles/c4ge9kggp1go. Accessed on 7 June 2025. Hargie, O., Dickson, D., & Tourish, D. (2004). Communication skills for effective management. Springer. Keszthelyi, C. (2025). AI in the classroom: a tool for learning or a shortcut to superficiality?. THINK magazine. University of Malta. Available at https://www.um.edu.mt/library/oar/bitstream/123456789/136032/1/THI NK47-AI.pdf Accessed on 21 June 2025. Marr, B. ( 2023, July 24) The difference between Generative AI and traditional AI: An easy explanation for everyone. Forbes. Available at https://www.forbes.com/sites/bernardmarr/2023/07/24/the-differencebetween-generative-ai-and-traditional-ai-an-easy-explanation-for-anyone/. Accessed on 22 June 2025. Nelson, P., & Weatherald, C. (2014). Cracking the code—An approach to developing professional writing skills. Social work education , 33 (1), 105-120. Nilson, L. B. (2023). Infusing critical thinking into your course: A concrete, practical approach. Routledge. Perkel, S. (2025, June 2025) 3 teachers tell us the changes they’re making in the classroom to address student’s rampant use of AI. Business Insider. Available at https://www.businessinsider.com/teachers-change-curriculumassignments-student-ai-cheating2025 6. Accessed on 21 June 2025. Sage (n.d.) Assistive and generative AI guidelines for authors. Available at https://www.sagepub.com/about/policies/ai-author-guidelines. Accessed on 5 June 2025 Southworth, J., Migliaccio, K., Glover, J., Glover, J. N., Reed, D., McCarty, C., Brendemuhl, J. & Thomas, A. (2023). Developing a model for AI Across the curriculum: Transforming the higher education landscape via innovation in AI literacy. Computers and Education: Artificial Intelligence , 4 , 100127. Vrågård, J., Brorsson, F., & Aghaee, N. (2024, October). Generative AI in Higher Education: Educators’ Perspectives on Academic Learning and Integrity. In Proceedings of The 23rd European Conference on e-Learning. Academic Conferences International. Wise, K. (2005). The importance of writing skills. Public Relations Quarterly , 50 (2), 37. Wise, B., Emerson, L., Van Luyn, A., Dyson, B., Bjork, C., & Thomas, S. E. (2024). A scholarly dialogue: Writing scholarship, authorship, academic integrity and the challenges of AI. Higher Education Research and Development , 43(3), 578–590. University of Twente (2024, June) Artificial Intelligence and Assessment (at the UT). Available at https://www.utwente.nl/en/examination/Additional_topics/AI_Assessme nt/ (accessed on 21 June 2025)