AI evidence extraction

Artificial Intelligence in Assessment: A Bibliometric Review of Research Development, Thematic Patterns and Research Clusters

Ashmimi Maisara Asha’ari, Anis Diyana Halim, Ketang Wiyono · International Journal of Instruction · 2026

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

8/10

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/10.29333/iji.2026.19238a

Methodology & findings

Study design

Bibliometric analysis of 68 Scopus-indexed publications from 2011 to July 2025, including citation analysis, h-index and g-index calculation, keyword co-occurrence analysis, and thematic mapping to identify research clusters and trends..

Main result

The study found that "The dataset recorded 1,581 total citations, with an h-index of 22 and g-index of 38, indicating steady growth that intensified after 2021 as AI adoption expanded in post-pandemic education." Additionally, "Keyword co-occurrence analysis revealed three dominant thematic clusters: technology-enhanced learning and automated feedback, machine learning with ethical considerations, and AI-supported formative assessment in science and engineering education."

Research paradigm

Positivist/Empiricist (bibliometric analysis of quantitative publication data)

Author conclusions

The authors conclude that "this study provides a comprehensive overview of global trends, leading contributors, and emerging themes that inform future research and policy directions in AI-driven educational assessment." They further note that "the findings highlight AI's transformative potential for adaptive, feedback-oriented learning" while acknowledging persistent implementation challenges.

Risk of bias

Database selection bias: limited to Scopus-indexed publications, potentially excluding non-indexed journals and grey literature; Language bias: likely English-language publications predominate in Scopus; Publication bias: only published work included, not unpublished or negative results; Geographic bias: overrepresentation of English-speaking countries (US, UK, Germany); Database selection bias (Scopus-indexed only, excludes grey literature); Publication bias (indexed publications may skew toward positive findings); Language bias (likely English-language publications); Database selection bias (Scopus-only), publication bias (peer-reviewed sources only), language bias (likely English-language emphasis), temporal bias (coverage through July 2025).

Open questions raised

Future research directions include: (1) addressing accessibility and equity in AI-based assessment systems; (2) improving curricular integration of AI tools in educational contexts; (3) enhancing teacher readiness and professional development for AI-driven assessment; (4) establishing ethical governance frameworks for AI in education; (5) strengthening interdisciplinary collaboration between AI researchers and education practitioners.
Future research should address challenges in accessibility, curricular integration, teacher readiness, and ethical governance of AI in educational assessment.
Future research needs to address accessibility, curricular integration, teacher readiness, and ethical governance in AI-driven educational assessment. The review identifies interdisciplinary research opportunities in automated feedback, machine learning applications, and data-driven evaluation in educational contexts.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 74%

Explore related topics

Evaluation & Benchmarks AI Governance Meta-Analysis