ChatGPT in Doctoral Supervision: Proposing a Tripartite Mentoring Model for AI-Assisted Academic Guidance

Aim/Purpose: The potential of Generative AI in education is expanding, yet its role in PhD mentoring and academic guidance remains underexplored. This study evaluates how ChatGPT-generated recommendations can support PhD research, particularly in fostering sustainable and resource-efficient doctoral education. Background: This study examines the ability of ChatGPT to provide structured guidance and actionable insights in PhD supervision. Using a real-world case study on disaster risk management, the research evaluates AI-generated recommendations across different prompt structures to determine their relevance, depth, and applicability to doctoral research. Methodology: A structured evaluation was conducted using input prompts with varying contextual details, including naive prompts, supervisor-selected keywords, ChatGPT-generated keywords, and topic-specific concepts. Five external academic experts assessed the AI-generated outputs for appropriateness, interrater agreement, and research alignment. Contribution: The study demonstrates that ChatGPT can enhance PhD supervision by providing structured academic recommendations, reducing administrative burdens on supervisors, and contributing to the evolution of a “tripartite mentoring model” where AI, supervisors, and students collaborate to tackle complex research challenges. Findings: AI-generated recommendations were most effective when structured around topic-specific concepts. Naive prompts also yielded relevant outputs, whereas keyword-based prompts resulted in less cohesive recommendations. Tailored prompts aligning with specific research pathways were rated as highly actionable and contextually grounded. ChatGPT demonstrated the ability to refine research methodologies and improve knowledge discovery. Recommendations for Practitioners: Universities may consider incorporating AI tools such as ChatGPT to support PhD supervision, particularly to provide structured feedback and guidance. Supervisors should explore AI-assisted mentoring to optimize time-intensive advisory tasks and enhance research productivity. Recommendation for Researchers: Researchers should explore the effectiveness of AI-driven academic guidance across various disciplines, assessing its impact on research quality, methodological rigor, and doctoral student development. Future studies may also investigate the ethical considerations of AI in PhD supervision, including potential biases in AI-generated recommendations and risks related to over-reliance on automated feedback. Impact on Society: By reducing supervisory workload and enhancing research efficiency, AI-driven academic guidance can promote equitable access to high-quality doctoral education, fostering innovation and sustainable educational practices globally. Future Research: Future research should evaluate AI-driven mentoring across multiple academic disciplines and multilingual contexts to better assess generalizability. Additionally, studies should explore ethical implications, including how disciplinary norms, cultural expectations, and linguistic diversity influence the effectiveness and appropriateness of AI in doctoral supervision.