Description
Seminar for PG research students in the School of EducationThis is a 90-minute workshop session, intended to support postgraduate research students with making effective use of generative AI in early and middle stages of research. Our emphasis is on practical uses of AI, although we discuss some of the broader considerations such as the problems of bias and hallucination. Throughout, we focus on retrieval augmented generation (RAG), moving beyond a single ‘zero-shot’, to providing the language model with source text from which it draws. We focus on the University’s Microsoft Copilot tool, but do not limit ourselves to this, making use of other platforms such as Google’s Notebook LM and the ‘deep research’ capabilities of Perplexity and Chat GPT. The approach is in-line with the University’s current policies around AI use, and we emphasise the need for criticality and academic integrity throughout.
Participants are introduced to the Google Scholar PDF viewer, and experiment with this to consider research papers in the context of the broader field. We introduce participants to how Copilot can be used to summarise and critique a research paper, and demonstrate how participants can enter into a dialogue with the chatbot to deepen their understanding of a text and related work. We introduce participants to the additional capabilities of Notebook LM, which extends the above to a corpus of related texts. We further demonstrate how ‘deep research’ models can provide a summary of a field. We discuss how participants could effectively use tools such as these to support their learning and to deepen their thinking. We discuss how these tools can be applied to participants’ own work, and introduce the University’s Write Now! Custom GPT for feedback on assignment drafts.
We demonstrate how speech-to-text tool in Office 365 offers quite reliable transcription. We explore how Copilot can be used to support qualitative analysis, focussing on applying Braun and Clarke’s method to a single transcript, and consider how researchers can or should make use of this technology to support them in their work.
We conclude with a brief discussion of the issues generative AI raises for academic integrity, and the unique contribution of human researchers to social science.
Estimated audience numbers (if applicable)
10| Period | 21 May 2025 |
|---|---|
| Held at | Centre for Educational Research |
| Degree of Recognition | Local |