Interpreting data from the use of AI in assessment and feedback: A student research perspective

Suh Colins is a student in the Faculty of Business who completed a Student Research Experience Placement (SREP) with LITE in Summer 2026. In this blog, he reflects on the research he conducted about student and staff attitudes towards AI use in teaching and learning.

Artificial Intelligence has gone through a significant expansion in the past few years and is affecting almost every part of our daily lives. It is now a principal challenge in academia as universities try to encourage the ethical use of AI while still upholding their core values of academic integrity, critical thinking, and creativity. The question remains as to what extent staff and students should be allowed to use AI in both their teaching and learning processes. This discussion has moved into the classroom through an institutional research project at LITE that investigates staff and students’ acceptance of using AI in teaching and learning.

As a student researcher, I analysed semi-structured staff and student interview transcripts, developed codes, identified patterns, and examined differences in responses across participants and demographic groups in relation to the research questions and objectives. This blog discusses how I managed my positionality, engaged with post-plagiarism concepts, and navigated challenges for qualitative researchers.

Positionality and Analysis

This project prompted me to reflect critically on my positionality. I entered the project assuming that everyone is using AI, especially university lecturers. This was shaped by my previous work experience in HR and my current PhD research on AI adoption and the Future of Work. As an active AI user, I used AI tools in my previous workplace in areas such as recruitment, performance management and learning and development. This experience had influenced my expectation that university lecturers would generally be experts in using AI in their lesson preparations, teaching and feedback, and would actively support students in using it in their learning process in preparation for what they will encounter in their future workplace. However, these assumptions were challenged when I started engaging with the interview data of staff and students on their use of AI.

I read accounts of lecturers who are not using AI at all, while others had limited knowledge of the range of AI tools available for teaching and learning. This made me more conscious of the need to distinguish participants’ experiences from my own expectations and to avoid imposing my understanding of AI and other matters related to the research topic onto the data. I had to read the transcripts carefully, question my initial interpretations, and allow the accounts of participants to guide the analysis.  This was the most important part of my reflexive learning during the project.

The timing was particularly relevant, as I was preparing to begin data collection for my PhD, allowing me to apply my existing knowledge of qualitative research to practical interview data. After engaging with literature on thematic analysis, including Braun and Clarke (2000), Evans and Lewis (2018) and Christodoulou (2024), as well as coding by Boyatzis (1998) and Flick (2013), this project gave me an opportunity to develop these methodological skills through practical work. I saw coding as a technical process of attaching labels to sections of a text but realised the difficulty in practice is deciding on what to code, how many words to add to a code, and what meaning a participant was conveying. I also learned that codes are not fixed from the beginning of the analysis. As I got familiar with the transcripts, some initial codes became too broad. Others needed to be merged or separated, while others needed to be renamed. All of this became clearer as I continued to engage with the context of the interviews.

Post -plagiarism

One particularly interesting aspect was learning about post-plagiarism: a conceptual framework developed by Sarah Elaine Eaton (2023) which defines an era in which historical notions of what it means to write and create ethically are being challenged. Post-plagiarism challenges traditional ways of thinking about plagiarism by encouraging the use of AI, even though some scholars describe large language model (LLM) outputs as nothing more than high-tech plagiarism (Chomsky et al., 2023). Its six tenets affirm:

• Hybrid human-AI writing will become normal,
• Human creativity will be enhanced,
• Language barriers will disappear
• Humans can relinquish control but not responsibility,
• Attribution will remain important, and
• The historical definition of plagiarism no longer applies.

What surprised me was the divergent opinions some staff and students have about these tenets. Some staff see it as defeating the purpose of university education and academic integrity: unethical, no critical thinking, no creativity, no originality or authorship, and a system that encourages laziness in the learning process. Some saw possibilities for more flexible and creative approaches to learning and the need to redesign the assessment process to integrate this new technology.

These contrasting views changed my perspective on AI in education. I realised how technology can challenge existing educational practices of teaching and learning. Rather than controlling or stopping staff and students’ use of generative AI, I feel this is an opportunity to rethink the assessment process by making it more innovative and creative: what is being assessed and how it is assessed. This can be done through presentations, projects, practical tasks, discussions, or workplace-based activities. The aim is to test critical thinking, reasoning, reflection, and problem-solving skills.

The interviews also revealed that attitudes towards AI are much more complicated than simply being “for” or “against” the use of generative AI. One of the most curious things I discovered was that some staff categorically refused to use any form of AI because of environmental concerns and the capitalist agenda behind it. Their position was not based primarily on academic integrity or fear of technology, but on the environmental costs (energy and water consumption, air and water pollution) associated with AI. This point got me thinking about the harm AI is really causing to the environment and the need for policies to be put in place to regulate the development and use of AI systems.

Working with my supervisors, Hanlin and Priscila, was another important part of the placement. During our first meeting, I was nervous and did not really know what was expected of me because it was my first time engaging in data analysis in a project where I was not involved in the design. Unlike my PhD project, where I am the master of my concepts, data collection and analysis, and my supervisors basically guide me through these stages, I was entering an ongoing project. Their guidance helped me understand that good research involves collaboration and continuous learning, and they helped me to understand and engage with data at the analysis stage, though I was not part of the design and data collection process.

At the end of my placement, I learnt to accept advice and corrections, pay close attention to detail, communicate my ideas clearly, and remain open to interpretations different from my own. I also developed greater discipline in meeting deadlines, reporting work done, and maintaining consistency in the analysis. These may appear to be simple research skills, but experiencing them within an active project made their importance much clearer to me.

From my experience in this project, if you are doing or planning to engage in qualitative research, then you need to remember the following: Do not let your previous knowledge or assumptions about a subject affect your analysis. You should be transparent about your positionality while interacting, analysing and presenting your data. Also, qualitative research is time-consuming. Plan sufficient time for data analysis. Do not underestimate the time needed to transcribe interviews, develop codes, identify patterns, and move from description to interpretation, especially when you have rich interview data. These lessons will certainly help me with my own PhD data analysis process.

---

Interested in the topics considered in this blog? Listen to our podcast episode featuring Professor Sarah Elaine Eaton from the University of Calgary and Dr Beatriz Moya-Figueroa from the Pontificia Universidad Católica de Chile, where they discuss post-plagiarism, authorship and attribution in the age of AI with LITE Higher Education Research Support Officers Hanlin Zhang and Priscila Riffo Salgado.