If you’ve found yourself wondering how much Generative AI is reshaping student academic writing – and what on earth to do about it – you’re not alone. Across UTS, academics are grappling with the same questions:
- How is GenAI changing the writing students produce?
- How much should we encourage the use of GenAI and in which stage(s) of writing?
- And how should we teach and assess writing now?
These questions motivated the Academic Language and Learning (ALL) team to conduct an exploratory study across all UTS faculties in 2025, funded by a grant from the Association for Academic Language and Learning. This blog post shares some highlights from a journal article we have just published based on findings from a survey with 58 UTS academics across all faculties.
GenAI is reshaping student writing, but unevenly
Across the 58 UTS academics who responded to our survey, most reported that GenAI is influencing the way students write. Yet its perceived impact varies considerably between disciplines. Business, Transdisciplinary School, Engineering, IT and Science academics reported the strongest sense of change, with half or more of those respondents stating GenAI has influenced writing “a lot.” Health, Humanities (Faculty of Design and Society) and Law academics were less certain about the impact of GenAI on writing or reported a lower level of impact.
Despite these differences, 93% of academics across all faculties still believe writing is an important skill in the age of GenAI. So, even as AI tools draft, edit and restructure text, academics see writing as a necessary skill to develop. Why? Because writing is still a way of thinking, analysing and making sense of disciplinary knowledge.
Academics are adapting, but differently
Over 70% of respondents said they had already adapted assessment or teaching practices in response to GenAI. However, there was a lot of variation in how they have adapted, with these being some of the stand-out themes:
1. Critical thinking and contextualised tasks are ever important
Academics across all faculties are designing tasks that ask students to critique AI-generated text, reflect on how (and whether) they used AI, or draw on personal, professional or contextualised experiences that GenAI can’t easily replicate.
2. Assessment design is becoming a frontline defence
Several academics – particularly in Humanities, Law and Health – are redesigning assessment tasks to:
- reward originality and disciplinary thinking
- reduce the value of GenAI-generated generic responses
- require rigorous, verifiable referencing
- incorporate vivas, in-class writing or staged submissions
- explicitly teach ethical use of GenAI
3. Others are still unsure what to do next
A small group of academics reported they hadn’t yet adapted anything, often citing a lack of confidence, clear guidelines or understanding of GenAI’s capabilities and risks.
A notable disciplinary divide
One of the most striking findings was the clear disciplinary patterning of understandings, practices and adaptations:
- Engineering, IT & Science – more open to AI as a writing support tool, especially for drafting and editing technical text
- Health – variety of responses from asking students to engage with AI tools to reluctance to discuss AI explicitly with students; eager for more guidelines
- Humanities (Design & Society) – strongest concerns about critical originality, interpretive reasoning, and “thinking-through-writing”; some academics reported prohibiting GenAI use in certain tasks, but embracing it for image generation
- Law – already accustomed to oral assessment and professional guidelines, with a cautious but pragmatic embrace of AI
- Business & Transdisciplinary School – experimenting with GenAI integration, particularly in early stages of writing, and the use of experiential learning tasks to prevent AI use
These differences may not be arbitrary: they seem to map to the epistemologies of each discipline. In fields where writing documents thinking, interpretation, or judgement, it is essential that students still do the majority of those tasks to demonstrate learning, and GenAI needs to be integrated very carefully. Where writing is seen more as formalised communication of results, academics tend to be more relaxed about adopting AI tools for the drafting of reports.
Tailored strategies and stronger support
So what does this mean for teaching practice at UTS? Our research suggests 2 main takeaways:
- There’s no one-size-fits-all approach to GenAI and writing – disciplines need tailored strategies grounded in their writing practices, assessment cultures and epistemologies
- Academics need more support to grapple with emerging challenges – clear guidelines, professional learning and discipline-specific conversations are crucial
Read the full article here: Exploring disciplinary academics’ perspectives and practices regarding student academic writing in the age of GenAI (open access)