Writing as ‘passing’ and the role of generative AI by Helen Beetham

‘Generative AI’ or synthetic text models have shone a challenging light on assessment, particularly the assessment of student writing. Despite well-known flaws and frailties, synthetic text can easily be taken for human writing. In fact, it has surface features that student writers are often advised to follow, such as structured arguments and hyper-correct grammar. This new technology arrives into an economy that already rewards ‘passing’ a set of instrumental norms, and in which students are anxious to ‘pass’ as credible in their production of text. Synthetic text shows up these ongoing problems with writing for assessment, while introducing new ones. For example, what is legitimate and what is helpful in digital writing support? How to design assignments that integrate the use of synthetic tools or, alternatively, how to ‘proof’ against their use? How to value student writing at all, if generative AI can produce scholarly and professional text?