AI Assistant for PhD Supervision

Wynn, Martin G ORCID: https://orcid.org/0000-0001-7619-6079 and Skuridin, Alexander (2025) AI Assistant for PhD Supervision. In: 3rd Symposium on AI Opportunities and Challenges (SAIOC), May 13th 2025, Online. (Unpublished)

Preview

Slideshow (Slides presented at the conference May 13 2025) 15053 Wynn, M, and Skuridin A. (2025) AI Assistant for PhD Supervision.pdf - Presentation Available under License All Rights Reserved. Download (1MB) | Preview

Abstract

Technological breakthroughs in Artificial Intelligence (AI) are reshaping diverse industries, including higher education, where AI-agents hold considerable promise for enhancing learning outcomes and operational efficiency. Traditionally, higher education has been human-centred, relying on the expertise of lecturers, administrators, and academic leaders to develop curricula, manage resources, and uphold quality standards. However, AI agents which employ Large Language Models have the potential to instigate a significant shift towards AI-driven organisations. Algorithms can analyse large-scale student and organisation data, automate administrative tasks, and assist in designing personalised learning paths. AI supports rather than replaces human expertise - serving as a "co-pilot" to lecturers and administrators. Examples include automated course scheduling, quality assessment, transcribing and summarising meetings, and student helpdesk support. Yet questions remain about how to implement AI responsibly, maintain educational values, and manage the transition from human-led to AI-driven organisations. In this context, and drawing upon established project management and organisational methodologies, this research focuses on the role of the doctoral student supervisor and develops an AI agent to replicate and/or assist in the student supervision process. The research methodology comprises three phases: 1. Examination of Existing Practice (AS IS); 2. Transformation to AI-Assisted Processes; and 3. Validation and Measurement, this being both qualitative (e.g., feedback from the supervisor) and quantitative (e.g., efficiency metrics, time reduction). The technology deployment uses the Make.com platform to integrate ChatGTP with Google docs, gmail and other components (Cloud document converter and router). No programming skills were required to produce this AI Assistant that can receive a doctoral project proposal via email and provide a detailed assessment within minutes. Although this remains a pilot study based on the initial project proposal phase, it will be expanded to encompass the wider PhD supervision process. The study is significant in that it demonstrates the power of such Assistants and also the ease with which it has been assembled and tested. The research aims to formulate principles and a managerial method for employing AI agents that may act as a catalyst and worked example for application in other contexts within this, and other, universities.

Item Type:	Conference or Workshop Item (Speech)
Subjects:	T Technology > T Technology (General)
Divisions:	Schools and Research Institutes > School of Business, Computing and Social Sciences
Depositing User:	Martin Wynn
Date Deposited:	21 May 2025 13:54
Last Modified:	21 May 2025 14:15
URI:	https://eprints.glos.ac.uk/id/eprint/15053

University Staff: Request a correction | Repository Editors: Update this record