Commentary: AI has changed the way my students write and that worries me

Commentary: AI has changed the way my students write and that worries me The problem is when students begin to believe that having access to well-phrased answers is the same as developing genuine knowledge, says university lecturer Kennedy Albar. SINGAPORE: In recent semesters, something has changed in the way some students in my ethics and communications classes write. In the past, students who wrote well often wrote differently, some more direct, some conversational. Some took risks with unusual examples or arguments. Even weaker writers tended to have recognisable habits or ways of framing ideas. Increasingly, however, I find myself reading submissions that feel polished but interchangeable. Arguments arrive neatly packaged: a balanced introduction, three structured points, a counterargument followed by a conclusion that sounds thoughtful without saying very much. The same phrases emerge even when discussing entirely different topics. The answer, unsurprisingly, has been generative AI. To be clear, this is not an extensive issue. But even a handful of such cases worries me, especially given the recent announcement that students across Singapore’s institutes of higher learning will learn AI skills tailored to their fields of study from 2027. Students now have access to powerful tools. When they begin to believe that having access to well-phrased answers is the same as developing genuine knowledge and understanding, that will be the problem. Take the example of my student who submitted a piece with the terms “social responsibility” and “ethical implications”. When I asked what those terms meant in relation to his example, the student could only repeat the phrase rather than explain his reasoning. In the end, the student admitted to using AI to generate ideas and structure the piece, because “I didn’t know how to start”. Students sometimes submit carefully balanced arguments on controversial issues, but struggle when asked which side they personally find more convincing or why. Other times, they stumble when asked why they chose a particular example or source, before admitting that it was because it “fit the argument” rather than because they fully understood it. In the past, these would have been revealed with weak writing. The argument might be underdeveloped or students might struggle to connect the evidence to the claim. These showed us educators where the thinking had broken down. Now, AI can cover over some of those cracks. WHY ORIGINAL WRITING MATTERS In many cases, students are not trying to deceive anyone. Perhaps they are unsure how to begin, or they use it to improve their language or organise ideas they already have. Used thoughtfully, AI can be a valuable learning aid. But has the student made a genuine choice in the argument? Can they defend their claim when challenged? Can they restate their argument in their own words? These questions matter because a piece of writing is not the final product. In higher education, writing is one of the ways students learn to think, notice what is important, organise complexity, weigh competing ideas and develop judgment. Students need to experience the discomfort of thinking without immediate assistance, a process educational psychologists call the productive struggle. Because that discomfort is often where learning begins. Already, research suggests that while AI can improve writing fluency and efficiency of written outputs, heavy reliance on these tools may also reduce opportunities for deeper engagement with ideas, independent problem-solving and the kind of cognitive effort that supports deeper learning. This has implications beyond the classroom when students graduate into the workforce. As someone who teaches ethics and communication modules, I also think there is an ethical dimension to this conversation. Using AI is not inherently unethical. The more difficult question is what happens when students present arguments or ideas as entirely their own without meaningfully engaging with them. At its core, education is built on intellectual responsibility: not simply producing answers but taking ownership of how those answers came to be. TIME TO GO BACK TO BASICS Yet, I do not think the answer is to pretend AI does not exist or to ban it entirely. Nor do I think educators should spend all their energy trying to create “AI-proof” assignments. Students will graduate into workplaces where AI literacy is increasingly expected so they will need to learn how to use these emerging tools responsibly and critically. At the same time, though, traditional practices such as in-class writing and asking students to explain their arguments orally may become more important in an AI-rich environment. In my own teaching, I have begun to place greater emphasis on these practices that make thinking visible. These are not dramatic innovations. If anything, they are old-fashioned. But they reveal what a polished final submission sometimes cannot. I would rather read a piece of writing that is imperfect but genuinely thoughtful. What educators hope for students is that develop something deeper: the confidence to make judgments, defend ideas, take intellectual risks and recognise when an argument is truly their own. Kennedy Albar is a part-time lecturer in communication and ethics at the Nanyang Technological University (NTU). The opinions expressed are those of the writer and do not represent the views and opinions of NTU.