AI evidence extraction

Integrating Artificial Intelligence in Postgraduate Supervision: Emergent Opportunities, Challenges, and Strategic Responses for Institutions

Obrain Murire, Bramwell Kundishora Gavaza · Open Books and Proceedings · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

2/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.38140/obp4-2026-13

Methodology & findings

Study design

Systematic literature review (SLR) following PRISMA guidelines.

Sample

N = 37, 2 groups

Primary method

Qualitative thematic analysis guided by three research questions. The authors state: "This section presents the findings thematically, guided directly by the study's three research questions." No quantitative statistical methods were employed. Literature screening and selection used predefined inclusion/exclusion criteria applied systematically. Quality assessment used the CASP checklist applied independently by both authors with peer debriefing to enhance rigour and trustworthiness.

Main result

The study found that AI introduces significant improvements in three key supervisory domains: "administrative efficiency, academic writing and research support, and student engagement." Specifically, "AI tools substantially reduce supervisors' routine administrative workload" through applications such as "Grammarly, Turnitin, and AI-enabled dashboards" that "automate substantial portions of text analysis, error detection, progress tracking, and time management." Additionally, "AI significantly enhances students' academic writing and research capabilities" through tools like "Iris.ai, Connected Papers, and Research Rabbit" that provide "sophisticated literature discovery, semantic search, and citation network mapping." However, the study also identified that "AI introduces risks of algorithmic bias and unequal learning experiences" and that "extensive use of AI may erode critical thinking and analytical reasoning skills."

Reports effect sizes.

Research paradigm

Interpretivist/Qualitative

Author conclusions

The authors conclude that "While AI offers transformative potential for postgraduate supervision, its adoption must proceed with a critical, human-centred perspective." They emphasize that "Technology alone cannot address the relational, ethical, and pedagogical complexities inherent in research supervision. Thus, the future of supervision lies not in replacing the human mentor but in reimagining supervisory practices wherein AI becomes an ethical, equitable, and transparent partner in fostering research excellence." They further assert that "responsible AI adoption hinges not on technological capability but on institutional readiness, pedagogical integrity, and sustained human involvement. AI can enhance postgraduate supervision only when supervisors and institutions engage with it critically, reflexively, and ethically."

Risk of bias

Publication bias - only peer-reviewed literature included, no grey literature; Time-bound literature (2015-2024) may miss relevant earlier work; Selection bias in inclusion/exclusion criteria application; Language bias - database searches likely English-language focused; Rapid technological change renders findings potentially outdated; Limited diversity in sample characteristics across reviewed studies; Selection bias: literature limited to English-language peer-reviewed sources between 2015-2024; Publication bias: only published studies included, unpublished/gray literature excluded; Recency bias: rapid AI technological advancement may render findings quickly outdated; Sample limitation: constrained by size and diversity of available samples in reviewed literature; Methodological bias: PRISMA approach limited capacity to capture emergent practice-based themes; Selection bias: Only peer-reviewed literature from 2015-2024 was included, potentially excluding grey literature, preprints, or older foundational works; Publication bias: Systematic review relies on published studies which may overrepresent positive findings; Database coverage bias: Search limited to four databases (Scopus, Web of Science, Google Scholar, ERIC), potentially missing relevant studies in other databases; Language bias: No mention of language restrictions, but English-language bias likely given database selection; Recency bias: Focus on 2015-2024 period may not capture earlier foundational literature on supervision; Author perspective bias: Two-author review team may have shared blind spots; limited description of inter-rater reliability beyond CASP checklist application

Limitations

The authors state that "While the PRISMA method offered rigour in identifying relevant literature, it constrained the study's capacity to explore emergent, practice-based themes that a grounded theory approach might have captured." Additionally, "the absence of a longitudinal design similarly limits our ability to detect the evolving impacts of AI over time." They further note that "Given the rapid nature of advancements in AI technologies, some findings may become outdated quickly." The study's "scope of the reviewed studies was also constrained by the size and diversity of available samples, which affects the generalisability of the conclusions."

Open questions raised

Future studies should involve larger, more diverse samples and employ mixed methods to enhance generalisability
Longitudinal studies would facilitate tracking the continued impact of AI on postgraduate supervision
Discipline-specific studies might uncover the adoption and experience of AI across various academic fields
Further research into the ethical, psychological, and institutional dimensions of AI adoption is required
Need for empirical, practice-based research on AI integration outcomes in real supervision contexts
Absence of longitudinal design limiting insights into evolving impacts of AI over time

Extracted from: pdfAgreement 63%

Explore related topics

AI Literacy Training Doctoral Training Literature Review