Image generated with Flux (Pro)

Context

Dissertations and theses are closely related to research, where the use of AI tools typically extends beyond just the formal writing of the final dissertation or thesis, but also includes students integrating AI tools throughout the research process. The following examples, drawn from conversations with students, highlight some of the ways students incorporate AI into their dissertations and theses across honours, master's and PhD.

A recurring sentiment that has emerged from conversations with students is that they are cautious when using AI tools. Many students express that while they don’t fully trust AI to handle the specific details of their research accurately, they do find it useful for general background information, initial explanations, or elaborating on content they’ve already created. Rather than relying on AI for authoritative answers, they see it as a supportive tool to clarify ideas, enhance expression, or explore different ways of presenting their knowledge.

Ideation

Generating new research questions: Students use AI tools as brainstorming partners to develop new research questions. AI can assist by identifying emerging trends, offering varied perspectives, suggesting specialised angles, or revealing gaps in current research.

Reported examples of tools used: ChatGPT, Claude, Copilot

Literature Review

General understanding: Students use AI tools to build foundational knowledge or clarify broad concepts, especially when they are entering a new field or starting unfamiliar research.
Summarising research papers: Students use AI tools to efficiently summarise research papers. They may use an AI tool to summarise a single article, or they may input a collection of papers and ask the tool to identify key themes across them.
Organising files: Students use AI tools to sort research papers into themed folders. By uploading their documents, they receive a categorised list of filenames grouped by topic.

Reported examples of tools used: NotebookLM, ChatGPT, Perplexity, Copilot

Code Development

Writing code: Many coding platforms now offer AI plugins with code completion features. Students use these tools to automatically finish functions, correct syntax errors, and generate code documentation.
Debugging: When students run into bugs in their code, they feed the code into AI tools, which assist them in identifying and fixing the errors.
Data visualisation: Students input data into AI tools to receive suggestions on possible visualisations and to identify potential trends within the dataset. In addition, these tools can generate code to create the recommended visualisations.
Fine-tuning codes: Students use AI coding suggestions to suggest finer level codes, and show the evidence for the proposed codes.

Reported examples of tools used: GitHub Copilot (and GitHub Copilot integrated with VS code), ChatGPT, Windsurf, NVivo 15

Image Generation

Creating images: Students use AI tools to create images that support and communicate their research. While these visuals are most commonly generated for presentations, they are sometimes also included in dissertations or theses.

Reported examples of tools used: Midjourney, Canva AI, ChatGPT

Embedded Systems Configuration

Installation and setup: Hardware and software often involve complex manuals. Students using new equipment or integrating different systems turn to AI tools to help them interpret installation instructions and troubleshoot error messages.

Reported examples of tools used: Google Gemini

Research interviews

Preparing for conversations: When preparing to interview research participants or engage with stakeholders, students often use AI tools to simulate mock conversations. This helps them anticipate the natural flow of dialogue, assess whether their question sequence makes sense, and build confidence in their approach. AI tools can also assist in brainstorming new questions to include.

Reported examples of tools used: ChatGPT, Claude

Qualitative data analysis

Coding and theme identification: Students use AI tools to analyse interview transcripts, survey responses, and field notes. AI helps identify recurring themes, suggest coding frameworks, and highlight patterns across large datasets.
Transcript processing: Students input transcripts into AI tools to categorise responses, extract key quotes, and identify emotional tones within participant data.

Reported examples of tools used: NVivo 15 (with AI-assisted coding), ChatGPT, Claude

Writing

Refining and editing text: Students use AI tools to refine their writing by correcting grammar, suggesting alternative phrasing, and rewriting text in a more professional tone. They apply these tools not only to their dissertations and theses but also to grant applications, emails to collaborators, and conference talk abstracts.
Expanding bullet points: During the dissertation or thesis writing process, some students input bullet points into an AI tool and use it to expand those ideas into full paragraphs or sections.
Citations: Students use AI tools to format their references according to specific citation styles and to organise them alphabetically.

Reported examples of tools used: Grammarly (and Grammarly integrated into Overleaf), Microsoft Copilot, ChatGPT, Obsidian, Claude

---

Collective perspectives from HDR Supervisors and HDR Students on AI integration in their projects

School of Education and Social Work/ School of Art, Communication and English

Supervisors

The supervisors demonstrate a balanced and cautious approach to AI integration in HDR supervision. Their perspectives focus on the need for clearer institutional guidance to navigate AI use while maintaining academic integrity and protecting student data privacy. They view AI as an inevitable part of academic workflows, but advocate for thoughtful, guided implementation, rather than unrestricted adoption. They support formalising AI discussions through structured guidelines to benefit all stakeholders in the supervision process, moving beyond the current informal approach to create more systematic frameworks for AI use in HDR contexts.

Current practice:

Relies on existing university guidelines and shares them with PhD students
References resources like FASS HDR Supervision Forum materials
Emphasises secure AI platforms (such as Copilot) over potentially less secure alternatives
Conducts informal discussions about AI benefits and risks with co-supervisors and students
Maintains focus on confidentiality and data protection

Key concerns:

Establishing appropriate boundaries for AI use in supervision
Upholding academic integrity standards
Protecting sensitive student data
Need for formal, accessible guidelines rather than ad-hoc approaches

HDR students

They generally recognise AI's potential value while maintaining thoughtful caution about its appropriate application in doctoral research. They would like to have more open, and if possible, collaborative, conversations with their supervisors about appropriate boundaries for the project, alignment with institutional guidelines. They favour a principled approach that preserves research integrity while allowing for technological innovation in their scholarly practice.

Current Practice:

Selectively use AI for routine tasks like literature search refinement and formatting
Experiment with AI for brainstorming and concept mapping in early research phases
Seek supervisor guidance, where possible, before using AI for substantive research activities
Do their best to maintain transparency about AI use in research processes
Prefer established and personally suitable platforms with clear privacy policies, such as Claude.