“Those things I told you were just [false]statements. Isn’t it I was representing girls who live like that?” –

Participant X  

I had just finished data analysis and was about to write up my findings for my study, which explored the education experiences of six girls in Sub-Saharan Africa (SSA) who were characterised as ‘marginalised’. These girls had been recruited onto a vocational education and training programme by an international organisation, which sought to improve their lives through training them in skills for their livelihood. The findings were, to me at the time, a compilation of a neat and compelling narrative about these girls’ lives, which resonated a lot with the broader literature on girls’ education – mostly how vulnerable they are, how they lack access to important resources and how they lack voice and agency. I drew this narrative from these girls’ stories told in their own voices, a very important part of my methodology in this study. Nothing prepared me for the moment when one of the participants in my study (Participant X in the excerpt above) revealed to me in a WhatsApp message that all she had told me during my two long interviews with her was a misrepresentation and her narrative about being a ‘marginalised’ girl was not true!  

To say I was shocked and confused is an understatement. In that state of panic, I went back to the data, carefully observing my conversation with her and all the girls in this study, and to my horror, I realised that the stories of five out of six girls had inconsistencies and contradictions. At that point, disappointment set in, and I felt disappointed with myself, the participants, and my study. How did I miss that?  

My claim to having taken a constructivist stance in this study was tested by this revelation. I had gleaned from authors like Guba (1990) and others that constructivism denotes a relativist ontology, whereby there are multiple realities that are socially based and local and depend on the one who holds them. In this instance, my approach as a researcher was to (subjectively) interpret the participants’ experiences and perspectives, thereby taking an interpretivist epistemology (Cohen et al., 2007; Creswell, 2007; Crotty, 1998). What I had not anticipated was whether it matters if what the participants report is true or not and if the intention is to tell their story through their own voices, how this can be achieved when the participants have misrepresented themselves. 

After I recovered from my initial shock, I discovered that there were many important lessons from this experience, lessons which have shaped me as a researcher. For now, I will share this one: it is important for us as researchers to remember that the interview process is not neutral, and this has far-reaching consequences than we are consciously aware of. There are power dynamics at play, which are negotiated by the interviewer and interviewee both during and after the interview. For instance, as posited by others like Kvale (2006), typically, it is the interviewer who rules the interview, where the interview is an instrumental, even manipulative dialogue, in which the researcher has a goal and an agenda and maintains the exclusive privilege to interpret and report what the participant has said. Inherently, the expectation is that the interviewee will tell the interviewer what they want to hear. However, less emphasised, and often assumed during research, is the power that the participant holds. Not only do participants have the power to consent or not to the study, to choose to respond to a question or not, but they also have the power to select and transform details of their lives to suit the narrative that they want the researcher to remain with (Gubrium and Holstein, 2011). In my case, the participants exercised their power by presenting the persona they wanted to present and telling me the stories that they wanted to tell me, whether the stories were misrepresented or not. Their power certainly impacted me during data analysis and how I interpreted and reported the findings. 

Now, to the golden question: when it comes to reporting the findings, does it matter whether what the participants said is true or not? I think this is subjective to the researcher and their goals. For instance, I realised in revisiting what it meant for me to take a constructivist approach that my goal as a researcher was not to find or tell the truth because truth is indeed relative. I discovered when I engaged this line of thinking that the contradictions and inconsistencies of the participants in my study told an even more powerful story (Watson, 2006) though different from what I initially thought. In her work about ‘hearing’ girls’ voices, Khoja-Moolji (2006) emphasised that it is important to be open to diverse narratives about their lives rather than to reduce their voices to a homogenous narrative. For me, the complexity that came from the unexpected discovery underscored one of my strongest contributions about ‘marginalised’ girls, as I interpreted and reported all the stories – the truth, the half-truths, the outright misrepresentation.  

Does it matter to you if what your participants tell you is true or not?  

blog post by Jane Doka, a Research Associate in WELS – Faculty of Well being, Education and Language Studies.

Dr. Jane Doka is a skilled researcher with expertise in Comparative and International Education, specializing in the areas of youth transitions, inclusion, and gender within educational contexts. Her work emphasises the nuanced experiences of adolescents, particularly girls/young women characterised as marginalised within development frameworks, with an extensive background in both applied research and practical project implementation in the Global South. Her current and past projects underscore her commitment to ethical, culturally responsive research and the development of inclusive education systems. Jane is a member of the Centre for the Study of Global Development (CSGD) and is part of the Faculty of Wellbeing, Education and Language Studies (open.ac.uk)

Artificial Intelligence (AI), notably Chat GPT, as a language model, can potentially be misused for plagiarism due to its ability to generate coherent and contextually relevant text. While it’s a powerful tool for various legitimate purposes, there is a risk that unethical users may employ it to produce content without proper attribution or originality.

This was my belief last summer when I flagged several final-year scripts for potential plagiarism for unethical use of Chat GPT. Jonathan, in his recent blog, refers to my explanation of why I suspected students’ use of AI because their work seemed too perfect.

By ‘perfect’, I meant, as I wrote in the paragraph above, ‘coherent and contextually relevant text’. Actually, I did not write that first paragraph (only!); rather, it was provided for me by Chat GPT. (I doubt I could have expressed so accurately the way I had felt about AI.)  So, have I committed plagiarism?

I become unsure when I turn for help to The Open University’s Plagiarism Policy:  Plagiarism is using, without acknowledgement, someone else’s words, ideas or work.   How far can we reasonably describe a robot as ‘someone else’? Was I unethical to use an expression I had commissioned framed? Would it have been less unethical if I had edited Chat GPT’s text, or acknowledged its use, or have supplemented the text Chat GPT provided with appropriately referenced academic sources – my students did?

Further, what of our writing is totally original in any case? Bakhtin ([1952-3]1994) tells us, “Each utterance is filled with the echoes and reverberations of other utterances” (page 291). We continually adapt and adopt snippets of text from elsewhere and present them as our own. It is acceptable to consult a dictionary, a thesaurus or a Google search to help us write that coherent and contextually relevant text. Jonathan, in his recent blog post, asks what all the fuss is about regarding AI, and I wonder, should we be making a fuss?

Jonathan cites “Can IT think?” by Philip Ball (2023), who argues that AI should be treated with great caution. and I have come across descriptions of widespread exploitation of AI  with dubious results, such as the use of a Chatbot as a therapist – but is employing AI to aid our academic writing unethical?

Returning to Chat GPT for inspiration, it continued to advise (or followed my instruction to do so) about the existence of Open AI, their research company, which states their belief that “AI should be an extension of individual human wills” – an extension, not a replacement, then for human endeavour. This approach seems to resonate with Simpson (2023), a clinical teaching fellow, who advocates reframing the way (medical) students think about AI “not as an academic shortcut but as more of a companion”.   I like the idea of “companion” – like a dictionary or thesaurus – I also appreciate the concept of “shortcut” as a contested one.

Might we ask, in our potentially fraught, busy, complex lives, why we should not look for shortcuts in our academic life besides our everyday existence? And how much of that endeavour that AI use shortcuts form part of a valued academic activity? It’s saving thinking and editing time and providing that springboard to develop discussion, as it has for me above. Daher (2023) in Will Chat GPT be the disrupter academia needs? seems to cautiously embrace AI as “the spark that will change education for the better“, a means to reframe what we value in academic writing and to turn our focus towards critically evaluating sources.

I do not understand that argument. Surely critical evaluation already forms a key part of being an academic. And I do value that time of thinking, crossing out, rewriting, checking and editing; it’s part of the process that makes writing my own. I’m not looking for shortcuts, and I don’t plan to continue to make significant use of AI in my own work. But I don’t now think using AI in academic writing is necessarily unethical, and how far it is plagiarism is a discussion we need to have.

Chat GPT finished the 200 words I’d requested with a bland reassurance:  Encouraging responsible AI use can help ensure that the technology benefits society positively without contributing to plagiarism issues. (Chat GPT)

AI, then, is just another tool in our digital repertoire, and, as Jonathan asked, What is all the fuss about? I am still not sure…

by Jane Cobb

I have been an Associate Lecturer at the Open University since 2002, tutoring mainly English Language modules.  I live in Stourbridge in the West Midlands with my husband, two adult children and three Romanian rescue (street) dogs.  My recent EdD and my current research interests concern the multiple perspectives around feedback practices around assessed writing in HE.  This is my first venture into blogging, and I am looking forward to this creative space, where colleagues can share, debate, and discuss issues arising around their research.

For me, it started one Friday in July (2023) when I joined a Teams meeting with three OU colleagues who organised the WELS PGR Blog. We got talking about Artificial Intelligence (AI). One of the three colleagues thought it was ‘great’ and told us that she uses AI all the time and has been reading articles to get more information. A second colleague, an OU AL, said that she finds many of her students are using ChatGPT in their assignments and that she is beginning to get a sense of what AI-generated writing sounds like; it is often a little bit too perfect, in her view. The third colleague contrasted this experience to her own as she finds that there are some students who have never heard of ChaptGPT or AI.  

At this point, I felt that I had to confess that while I had heard a lot in the media about the risks of AI and rather less about its benefits, I had never actually engaged with it. The colleague who likes AI gave me the link for ChatGPT.  

Once the Teams meeting was over, I signed up for ChatGPT. There, I was faced with the front page giving some examples of possible enquiries ranging from creative ideas for a birthday party to explaining quantum computing. The same screen also outlines the capabilities of Chat. GPT. It is ‘trained’ to “decline inappropriate requests”. It then admits that its limitations include ‘occasionally’ generating “incorrect information”, “harmful instructions or biased content”. 

It then invites you to “Send a message”.  

I do not know if it was these limitations that made me stare blankly at the send-a-message request. Wouldn’t it be awful if my first message was deemed inappropriate? Would I know if I was sent harmful instructions or incorrect or biased information?  

As the English Men’s Cricket team was engaged in a test match against Australia, I thought I would start by asking about the names of fielding positions which, beyond wicketkeeper and slips, I have a very tenuous grip on. Immediately, ChatGPT presented me with a list of these positions. Very impressive. 

Emboldened, it occurred to me that I had had a supervision session with one of my EdD students. I had been trying to suggest that what she had been talking about in her most recent piece of writing could be related to Foucualt’s ideas about micropower and the role of ‘examination’ in disciplining populations. So, I asked about this. ChatGPT responded, saying these sorts of ideas were implicit in much of Foucault’s work. So, I asked for some examples. One of these was Discipline and Punish (Foucault, 199). I have a copy of this, so I found an example of where Foucault talks about how examinations are used. ChapGPT thought I was quite right!  

Coincidentally, shortly afterwards, I read an article by Philip Ball, “Can IT think?” (Ball, 2023), which not only outlines how AI uses algorithms to scan “vast banks of online data”  but also discusses the arguments for and against a possible ‘robot apocalypse’ and the dangers of AI being used to increase cybercrime and terrorism. Ball concludes by siding with experts who suggest that AI should be treated in the same way as new drugs and licensed for public use only after careful testing. Ball concludes:  

“ … in seeking solutions, we are to some extent flying blind because we do not know what kinds of minds these machines have- and because, in the absence of that knowledge, our impulse is to presume that they are minds like ours. They are not. It is time to take machine psychology seriously.” (Ball, 2023, p. 33). 

This made me sit up and take notice. I had quite enjoyed what seemed like a pleasant chat with Chat GPT about Foucault and was thinking that this was not that different from interrogating other sources of information. In a way, I had moved through a cycle starting with ‘What’s all the fuss about?’ (I don’t know) to ‘What’s all the fuss about?’ (It’s just a quick way to access information.) to return to What’s all the fuss about? (Does anybody know?) 

Jonathan Hughes

Image: Photo by Aleksandar Pasaric from Pexels: https://www.pexels.com/photo/what-is-this-is-all-real-text-with-yellow-background-3280211/

Jonathan is a member of the PGR Blogger editorial team.  Lecturer (access and curriculum) at the Open University’s Centre for Inclusion and Collaborative Partnerships, where I am also the Academic Conduct Officer and Assessment Lead. I am the Academic Lead for the Open University Badging Project which is developing the first university badged open courses outside the USA. I have been working as skills lead, author and critical reader on Open University Health and Social Care modules.

My research interests include widening participation and learning in later life, as well as later life sexuality. I am the chair of the Association of Education and Ageing.

In June 2014, I became a Senior Fellow of the Higher Education Academy.