Originality and the use of Generative AI.

Originality as systematic reflective and reflexive thinking.

At this time of year, I’m marking MA dissertations and teaching. I feel for those students I give low marks to. As I mark their submissions and write my feedback, I’m always thinking: how can I describe the problems I see- so that what I say in the ‘drop down’ is clear to them? My feedback is more often about how a student can ‘put more of themselves into their submission’ and not to give up on their own ideas.

It seems to me that many international students have a different understanding of what an assessor is looking for in their paper. Moreover, where they resort to using generated text in their submissions, it is too often a last resort rather than a planned activity and because they believe it offers something better than their own ideas. Perhaps they believe that assessors are looking for demonstrations of systematic knowledge when in fact they are more interested in demonstrations of critical thinking, reflection and reflexivity. It is my view that too many post graduate students devalue important assessment criteria relating to originality.

Too many students give too much energy to summarising, paraphrasing and synthesising a large volume of notes but give too little space to their own ideas and reflective forms of writing. Consequently, their paragraphs become too descriptive, less discursive and less well aligned with their stated intentions. Where a student is more confident about what they know and remains true to their stated intentions, they are more likely to construct a better narrative in their paper. In other words, those students that ‘stay with the problem’ they have identified are more likely to share original thinking and less likely to drift into accidental plagiarism- commonly identified by assessors as weak claims, unjustified statements, weak paraphrasing and inappropriate citations.

It is in the context of ‘giving up’ and using Gen AI, that I think it is helpful to talk about originality- described here as systematic and reflexive thinking- as opposed to ideas of systematic knowledge in an area of research. We should begin by thinking of originality as where a student selects an idea, a problem, or issue for discussion. This is normally evident in the title and research question of an assignment. In a dissertation there is also good scope for originality in the design of the research. However, most of the marks awarded for originality in a paper should be for the author’s voice, in the construction of the sentences and paragraphs where the author is more reflective or more reflexive. I would like postgraduate students to think about ‘self’ more often. More than that, I would like students to picture the balance of their writing as a set of scales tipping between individuality and plagiarism. On one side there are active expressions of thinking and reflexivity- strong author voice- and on the other there are passive expressions of secondary knowledge and borrowed voice.

In this metaphor there are degrees of originality and plagiarism. It seems to me that that an author can add weight to their claims to originality by giving more space to their positionality (Holmes 2020). This is as true for writing an essay or for undertaking empirical research for a dissertation. In other words it is always better to start with what you know before researching what you don’t know. In this type of narrative the author can more easily offer a description of their thinking in relation to what they believe to be true and their reasons for engaging with different academic studies. However, where Generative AI is used as a search tool and as a co-author to identify themes and suggest ideas, the student must acknowledge their use of AI as part of their creative process. They should describe how they used AI tools and for what purposes, for example describing the prompts they wrote and the conversations they had with AI. In this way students can demonstrate originality in the use of Gen AI as a research tool and the scales of assessment can be tipped back from plagiarism to originality.