AI evidence extraction

The use of generative AI tools for students’ research support within higher education institutions' libraries: a systematic literature review

Mfowabo Maphosa, Cyril Tlomatsana · Information Research an international electronic journal · 2026

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

9/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.47989/ir31263101

Methodology & findings

Study design

Systematic literature review following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodological approach.

Main result

The study found that "research interest in the use of generative AI tools for student research support by academic libraries has been growing" with "64.4% of publications in 2024", and that "the seventy-one publications had been cited 657 times in 506 articles with an average of 9.25 citations per publication". Key findings also indicate that "students are using a small set of freely available generative AI tools, particularly ChatGPT" and that "academic librarians are redefining their roles, and those of their libraries, through adaptations of generative AI tools to improve their services", including "AI to automate indexing and cataloguing of information resources as well as 24/7 virtual reference chatbots".

Research paradigm

Interpretivist/Qualitative synthesis with descriptive quantitative analysis

Author conclusions

The authors conclude that "there is a growing research interest in the adaptation of these tools to enhance students' research activities within academic libraries" and that "Academic libraries are also highlighted to be slowly adopting generative AI tools to improve their information services, as well as students' information engagement activities." They further state that "AI and digital literacy are emerging as key skills for students to possess to thrive in an AI-dominated educational environment and evolving knowledge economy." The authors emphasize that "Academic libraries need to continue investing in training students about the ethical use of AI and digital and information literacy" and call for "synergistic universal frameworks in integrating generative AI to improve students' research activities worldwide."

Risk of bias

Database selection bias: Only Web of Science was used; Scopus and ERIC considered but excluded due to smaller outputs, limiting grey literature and practitioner-focused research; Language bias: Only English-language publications included (eighty-seven articles screened, sixteen excluded for non-English languages); Publication bias: Reliance on peer-reviewed, indexed publications may exclude recent or emerging studies; Geographic bias: Dominance of Global North research (China and USA), with minimal representation from Global South countries; Subject area bias: Information Science/Library Science dominates 45.1% of included publications, potentially skewing findings toward library-centric perspectives; Database selection bias: Only Web of Science searched; Scopus and ERIC outputs noted as 'relatively small' but not fully explored; Language bias: Only English-language publications included (87 articles screened, reduced to after language filtering); Publication bias: Indexing bias favoring high-quality peer-reviewed sources may exclude grey literature and practitioner publications; Geographic bias: Authors affiliated with China and USA dominate contributions; global South underrepresented; Subject area bias: Information Science/Library Science represents 45.1% of publications, potentially overrepresenting this perspective; Database selection bias: only Web of Science searched; Scopus and ERIC considered but excluded; Language bias: only English-language publications included; Publication bias: indexed sources only, excluding grey literature and preprints; Geographic bias: overrepresentation of Global North countries (China, USA) in authorship; Potential selection bias in screening process despite PRISMA framework

Limitations

The authors stated that "our study is limited to a systematic literature review based on Web of Science data
Therefore, it only reflects research published in indexed sources and not on grey literature and emerging studies that have not yet been indexed." Additionally, they noted that "the Web of Science database was chosen for its rigorous indexing standards and citation tracking capabilities
While this ensures the inclusion of high-quality studies, it may also limit the representation of emerging or practitioner-focused research that is not indexed in this database." The authors also identified that "the geographical representation of the author's affiliation with the research interest is not balanced
This is due to authors affiliated with China and the USA dominating the research contribution, while countries in the global South are marginally represented."

Open questions raised

Limited research specifically focused on generative AI tools within academic libraries for research support
Lack of validation on the effectiveness of proposed AI-LSICF and other AI literacy frameworks
Absence of discussion on how academic libraries bridge inequality between students using paid versus free AI tools
Need for research on long-term impacts of generative AI tools on students' research skills, cognitive development, and academic performance
Mapping the integration of digital and information literacy within HEIs' curricula as foundational skills for AI tool usage
Research from multiple databases to capture broader global scholarship, particularly from developing regions

Extracted from: pdfAgreement 83%

Explore related topics

Systematic Review Automation Literature Discovery