AI evidence extraction

The Role of Artificial Intelligence in Modern Academic Research: A Systematic Review of Contemporary Literature (2015–2025)

Bayu Kurniawan, Fitri Wahyuni, Erin Putri Sosang, Nabilla Endmal Putri Iryas · Suluh Jurnal Bimbingan dan Konseling · 2026

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

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.33084/suluh.v11i2.12402

Methodology & findings

Study design

Systematic Literature Review (SLR) following PRISMA guidelines.

Main result

The systematic review found that "AI has become a transformative force that simultaneously accelerates and disrupts the modern academic research ecosystem." Specifically, the review identified three major conclusions: "First, AI has been shown to enhance efficiency and accuracy in various academic processes, as reflected in the CWLA assessment system, which achieved a correlation of 0.88 with human raters. Second, AI also poses serious challenges to academic integrity and authenticity, as AI-generated texts are becoming increasingly difficult to distinguish from human writing for both machine detectors and human experts. Third, the impact of AI is neither neutral nor universal, but instead mediated by existing social, cultural, and institutional contexts."

Research paradigm

Interpretive/qualitative synthesis of heterogeneous empirical evidence

Author conclusions

The authors conclude that "artificial intelligence (AI) has become a transformative force that simultaneously accelerates and disrupts the modern academic research ecosystem, particularly in education." They state that "responses to AI cannot remain merely technical or regulatory, but must involve a broader systemic transformation encompassing the reform of academic assessment design, the development of integrated AI literacy curricula, and the establishment of consistent global ethical frameworks for governing AI use in academic environments." They emphasize that "AI's effects are neither neutral nor universal, but are strongly mediated by social, cultural, and institutional contexts."

Risk of bias

Single database limitation (ScienceDirect only) - potential publication bias and selection bias; Language bias (screening criteria mentions language as a factor, likely English-language bias); Temporal clustering - dramatic acceleration of publications post-2022 following ChatGPT emergence; Geographic bias - underrepresentation of Global South perspectives explicitly acknowledged; Study type heterogeneity - included studies use diverse methodologies (surveys, experiments, qualitative, bibliometric) limiting direct comparison; Publication timing bias - review includes very recent 2025 publications potentially before peer review validation; Single database search (ScienceDirect only) may miss relevant studies in other databases; Language bias (English-language publications likely overrepresented); Geographic bias (underrepresentation of Global South perspectives); Publication bias (peer-reviewed articles only; grey literature excluded); Temporal bias (focus on 2015-2025 may miss earlier foundational work); Selection bias (screening by predetermined criteria may exclude relevant studies); Single database search (ScienceDirect only) - limits comprehensiveness and may introduce database selection bias; Language bias - articles screened by language criteria (likely English-language bias); Geographic representation bias - limited Global South perspectives in included studies; Publication bias - only published peer-reviewed articles included; gray literature excluded; Temporal concentration bias - dramatic acceleration of ChatGPT-related publications from 2022 onward may skew findings toward recent technologies; Disciplinary bias - focus on education and higher education contexts may not represent broader AI research landscape

Limitations

The authors state that "this review is limited by its reliance on a single database and the still-limited representation of Global South perspectives in the selected literature." They recommend that "future research is recommended to expand database coverage and the representation of Global South contexts, as well as to develop longitudinal studies capable of capturing the evolving impact of AI on research quality and the sustained development of academic competence."

Open questions raised

Limited representation of Global South perspectives in current literature on AI in academic research
Need for longitudinal studies capturing the evolving impact of AI on research quality and sustained academic competence development
Insufficient empirical validation of rapidly developing AI tools
Theoretical fragmentation across disciplines
Need for expansion beyond single-database searches to capture broader literature
Lack of binding policies at most universities governing AI use in research

Data: ICNALE GRA dataset (referenced in CWLA study by Uchida & Negishi, 2025); COCA corpus (referenced in Uchida 2024 study on corpus linguistics); s2orc-9K dataset (referenced in Buscaldi et al. 2024 citation prediction study); Web of Science Core Collection (referenced in Gorraiz 2024 bibliometric analysis); Scopus database (referenced in multiple studies including Tomaszewski 2024)Extracted from: pdfAgreement 79%

Explore related topics

AI Literacy Training Multi-Domain Literature Review