By Jasper Pauwels, 9 April 2026
In March, SENSE’s UniSIG was happy to welcome back Nigel Harwood, professor in Applied Linguistics at the University of Sheffield, who had kindly agreed to share his thoughts on academic writing and proofreading in this brave new world full of generative AI. Some SENSE members will remember Nigel’s previous talk about his research on proofreading student texts in the UK.
Research confirms what we already know
Nigel shared the findings of recent empirical research regarding students’ use of AI during their studies. Unsurprisingly, all research points in the same direction: the vast majority of students are using AI and for all kinds of purposes, from information searching and grammar checking to creating entire essays with minimum input from themselves. In addition, few of the students surveyed said that they knew their university’s stance on AI in education, and even fewer seemed to care about the potential harmful effects of AI use on their learning, the environment or even on the value of their degree.
What the Russell Group says on AI
The Russell Group, an association of 24 top UK research universities to which the University of Sheffield belongs, produced an AI policy document which turned out to be very positive about AI. The group considers generative AI a fact of life and believes that students should be prepared for using AI in an ethical and responsible way in their future workplaces.
The AI policy document formulates the following five principles for education:
- Support students and staff to become AI literate.
- Equip staff to support and guide students’ use of AI.
- Incorporate ethical use of AI in teaching and learning.
- Uphold rigour and academic integrity.
- Share best AI practices.
On paper, this all sounds very promising. In reality, however, the policy leaves many important questions unanswered, including how Nigel, as a lecturer, is supposed to support his students and adapt his assessment methods to generative AI use. In fact, the policy does not provide specifics of what is to be considered ethical and responsible AI use. How academic staff members are to respond to unethical or unresponsible AI use remains equally unclear. Finally, Nigel also raised the valid question of whether universities should embrace such a rapidly evolving technology so quickly; maybe universities should take pause and do what universities do best, that is, thinking critically in order to reach a well-informed decision.
Nigel experienced firsthand how the rise of large language models (LLMs) resulted in his students, many of whom are not native speakers of English, suddenly handing in perfectly polished texts overnight, even though his course instructions clearly forbade the use of AI. Like many lecturers, Nigel is unsure how to deal with this kind of student behaviour, especially since the new AI policy does not provide helpful advice on fraud prevention. Returning to three-hour invigilated exams and introducing oral defences might be part of the solution, but it remains to be seen whether the Russell Group universities are willing to change their assessment methods so drastically in the light of their techno-optimism.
What about proofreading at universities?
After Nigel’s presentation, we discussed the implications for SENSE members as language professionals working in or for academia. At first glance, it would appear that proofreading by humans, or even academic writing as a skill, has now been rendered obsolete. A pessimist would argue that cheating and making your way through your degree with minimum effort has never been easier. It was argued that, thankfully, there are still plenty of good students who want to put in the work and actually learn something at university. One way to help those students is actually old-fashioned human proofreading.
Learning how to write well requires intellectual hard labour, and we agreed that professional proofreaders – or editors if you will – can help students develop their writing skills. Rather than fixing their mistakes for them – what any decent LLM can do nowadays – proofreaders can explain to the student why they made certain interventions or elaborate how the student can express their thoughts more precisely. This is not only instructive for the student but also showcases the added value of the language professional.
As one SENSE member eloquently put it, with educative proofreading, you can help students find their own voice. And I personally expect that finding your own voice as an academic author might make all the difference in a world full of uninspired AI articles. Some SENSE members confirmed that academic journals are facing a torrent of AI-generated submissions. Which also goes to show that taking the easy option is a more widespread problem in academia.
Although Nigel’s message was generally not an optimistic one, we found enough reasons during our meeting to believe that academic writing and proofreading is far from dead and buried.
|
Blog post by: Jasper Pauwels |
