Early career academics say artificial intelligence is intensifying existing problems around student writing, leaving both learners and their teachers uncertain about the purpose of written assessment, according to research published this month.
AI exposing weak foundations in academic writing
The study, published by the Staff and Educational Development Association and reported in Times Higher Education, draws on interviews with early career academics (ECAs) at UK universities. It found that the emergence of generative AI has made “longstanding tensions more visible” in how institutions support student writing, particularly when expectations are unstated and guidance varies between departments.
Interviewees described a “pervasive lack of confidence” among students when it comes to academic writing. That insecurity, the authors say, has been strengthened by students comparing their own work to AI-generated text and by easy access to automated essay production.
“AI has exacerbated students’ confusion about the purpose of written assessments,”
The paper’s lead author, Kiu Sum, a lecturer at London Metropolitan University, told Times Higher Education that ECAs often observed students treating writing chiefly as an assessment task. When AI can produce essays and reports quickly, students are more likely to see writing as a product to finish rather than a process that develops critical thinking, analysis and communication skills.
Foundational skills slipping from view
Participants in the study reported that the availability of AI tools is obscuring the need to develop core capabilities that precede writing: critical thinking, evaluation, synthesis and paraphrasing. The research concludes that ECAs do not so much struggle with the technology itself as with unclear responsibility for writing development across institutions.
That uncertainty, the authors argue, is amplified because expectations around writing are often implicit. Without explicit, consistent guidance, academic staff and students alike are left unsure about what written assignments are intended to achieve.
- Students are comparing their writing to AI output, which can deepen insecurity.
- Some ECAs report students increasingly view writing as a product rather than a learning process.
- The study links high reported rates of student AI use — described as “near universal” in some research, with figures as high as 95 per cent — to these changing attitudes.
What ECAs say they need
Early career academics interviewed for the study called for clearer institutional guidance and shared ownership of writing development. The paper highlights uneven support for ECAs, with gaps in mentoring and professional development that leave newer staff less prepared to navigate assessment challenges posed by AI.
Those interviewed said the combination of student uncertainty and inconsistent institutional policy has made it harder to address the educational purposes of assessment. As AI tools become more widely used, the study suggests institutions must be deliberate about making writing expectations explicit and about teaching the thinking skills that underpin effective academic writing.
| Finding | Implication |
|---|---|
| Students increasingly compare their writing to AI | Heightened insecurity; diminished perceived value of writing |
| Writing expectations often implicit | Confusion about purpose of assessment; inconsistent teaching |
The research adds to a growing body of evidence that AI is reshaping classroom practice and institutional priorities. For educators and policymakers, the study underlines the need for explicit curricula and assessment design that foregrounds the development of critical thinking and the cognitive processes that lead to strong writing.
As universities around the world wrestle with how to respond to student use of generative AI, the experiences of early career academics — who often deliver the bulk of teaching — will be central to any durable solution that seeks to preserve the pedagogical value of written assessment.
